Abstract The Erlian Basin is located in the Central Asia-Mongolian fold belt between the Siberian and Sino-Korean Cratons. It is a Mesozoic continental rift basin composed of 52 individual fault-depressions. The main phase of rifting took place during the Early Cretaceous when a series of fluvial-lacustrine sediments were deposited. Each depression forms an independent sedimentary system and behaves as an independent petroleum system. Hydrocarbon source rocks are found in the upper Arshan and lower Tengger Formations. These are mainly type II source rocks and are mainly located in oil generation window at the present day. A series of oilfields and commercial oil flows have been found in the basin, highlighting its good petroleum potential. Many of these oils are heavy. Six oil samples from the Anan and ten from the Jirgalangtu Depressions have been subjected to routine geochemical analytical techniques in order to evaluate the origins. The methods used include gas chromatography of the saturated and aromatic hydrocarbon fractions, gas-chromatography-mass-spectrometry of the saturated hydrocarbon fraction and stable carbon isotope analyses. The trace metal elements of the biodegraded oils from the Jirgalangtu Depression were also analysed by atomic absorption spectroscopy. Two types of heavy oils : primary and biodegraded were identified on the basis of these data. The former includes both immature and mature heavy oils. A filtering-and-spill process was proposed to explain the origin of primary mature heavy oils (or tar-mat) in the Anan Depression. The biodegraded oils from the Jirgalangtu Depression were ranked and classified in terms of the degree of biodegradation, using a series of geochemical parameters based on the gas chromatographic concentrations and biomarker fingerprints of gas-chromatography-mass-spectrometry of the saturated hydrocarbon fraction. The relationship between oil saturation and porosity indicates that the heavy oils in the Jirgalangtu Depression were biodegraded after they accumulated.