Diabetic retinopathy (DR) is one of today's main causes of blindness in numerous developed countries worldwide. The underlying pathogenesis of DR is complex and not well understood, thus impeding development of specific, effective treatment modalities. Consequently, the use of animal models of DR is of critical importance for investigating the pathogenesis of and treatment for DR. While rats and mice are the most commonly used animal models of DR, the zebrafish now appears to be a promising model. Nonhuman primates and humans have similar eye structures, and both can develop spontaneous diabetes mellitus (DM). Although various traditionally used animal models of DR undergo a number of pathological changes similar to those of human DR, several human variations, e.g. retinal neovascularization, cannot yet be fully mimicked in any existing animal model of DM. Since both the animal models and the methods chosen for inducing DR have great influence on experimental results, a clear understanding of available animal models is vital for planning an experimental design. In this review, we summarize the mechanisms, methodologies and pros and cons of the most commonly used animal models of DR.