A wide range of imaging studies provides growing support for the potential role of diffusion tensor imaging (DTI) in evaluating microstructural white matter integrity in Alzheimer disease (AD) and mild cognitive impairment (MCI). Our review aims to present DTI principles, post-processing and analysis frameworks and to report the results of particular studies. The distribution of AD-related white matter abnormalities is widely discussed in the light of deteriorated connectivity within certain tracts due to secondary white matter degeneration; primary alterations are also assumed to contribute to the pattern. The question whether it is more effective to assess the whole-brain diffusion or to directly concentrate on specific regions remains an interesting issue. Assessing white matter microstructure alterations, as evaluated by group-level differences of tensor-derived parameters, may be a promising neuroimaging tool for differential diagnosis between AD, MCI and other cognitive disorders, as well as being particularly helpful in the interpretation of underlying pathological processes.