Due to human-driven environmental changes, planet Earth has entered the new Anthropocene era, with major impact on biological diversity recognized as the sixth mass extinction period. The concepts of biodiversity and ecosystem services (ES) have risen to objectify and measure the human impacts on ecosystems and the many-fold contributions of ecosystems to human well-being. Among global terrestrial ecosystems, tropical forests are particularly important for the conservation of biodiversity and for the provision of ES. Agricultural conversion, logging, hunting, commercial poaching and over-harvesting lead to deforestation, degradation and defaunation of tropical forests, with highly variable consequences depending on many local factors. In Central Africa in particular, biodiversity and ES have been far less studied than in other tropical regions, despite the vital roles of these tropical forests in the livelihood of tens of millions of people in a context of high poverty. A better understanding of the determinants of biodiversity and ES in Central Africa is crucial for improving human well-being and the resilience of forest ecosystems. Despite the still relatively preserved tree cover across the region, biodiversity and ES may differ depending on forest land management and allocations. Therefore, the objective of this thesis is to assess the conservation value of tropical forests in southeastern Cameroon, as well as the supply of ES and use by local populations, in three contrasted forest allocations: a protected area, a Forest Stewardship Council (FSC)-certified logging concession, and three community forests. First, we assessed the conservation value of the three forest allocations, examining species richness and composition of two taxonomic groups: mammals inventoried with 44 camera traps, and dung beetles inventoried with 72 pitfall traps (Chapter 2). We also aimed to identify the determinants of forest conservation value, disentangling the effects of forest allocations, proximity to human settlements (villages and roads), and local forest habitat. Mammal and dung beetle species showed lower species richness in the community forests than in the protected area, and intermediate values in the logging concession. Proximity to human settlements and disturbance was negatively correlated to species richness of both groups, negatively correlated with species body size, and associated to the loss of the most threatened mammal species. The high species variability among forest allocations (i.e., spatial turnover) suggests that any conservation initiative should integrate many sites to protect a multitude of species, and not only large isolated areas. The high conservation value of the protected area has been confirmed, and the logging concession can play a complementary role in conservation strategies through landscape connectivity. In contrast, community forests are particularly defaunated due to their proximity to roads and villages, but they still provide wild proteins to local populations. Second, we assessed the perceptions of the supply of ES by tropical forests to local populations, and the determinants of these perceptions (Chapter 3). We evaluated the significance and abundance of ES by conducting a questionnaire survey with 225 forest stakeholders. The most significant ES perceptions were provisioning services (93% of respondents) and cultural services (68%), while regulating services were much less reported (16%). The perceptions of ES abundance were relatively homogeneous among forest allocations and respondents. Bushmeat provision has been identified as the only significant ES for local populations that is not supplied in high abundance. Third, we depicted the use of ES by local populations in three villages, and we evaluated its determinants and sustainability (Chapter 4). We used diverse interviews and field surveys to assess three provisioning services (bushmeat, firewood, and timber) and five cultural services (cultural heritage, inspiration, spiritual experience, recreation, and education). On average, local populations consumed 56 kg of bushmeat person–1 year–1 (hunting zones covering on average 213 km² per village), 1.17 m³ of firewood person–1 year–1 (collection zones on average 4 km² per village), and 0.03 m³ of timber person–1 year–1. On average, 59% of respondents recognized the importance of cultural services. The main determinants of ES use were forest allocations, population size, and deforestation rate, and we also showed slight differences between Baka and Bantu people in the use of cultural services. Firewood and timber have been shown to be used sustainably by local populations in this area, whereas bushmeat hunting and consumption have exceeded sustainability thresholds. Finally, the main findings of the thesis are summarized and their practical implications are discussed, in particular for the role of forest allocations (Chapter 5). The potential reconciliation between conservation and the sustainable use of tropical forests is discussed. Methodological feedbacks are given for the use of mammals and dung beetles as biodiversity indicators. Research perspectives are presented for a better understanding of the interactions between biodiversity and ES. Finally, different perspectives for integrating the concept of ES in tropical forest management are given: for instance, identifying and resolving conflicts among stakeholders, raising awareness, making decisions, or evaluating the effectiveness of conservation measures. In particular, ES are increasingly used in concrete management applications, such as FSC-certification, payments for environmental services, UNESCO Man and Biosphere Reserves, and various development projects.