Five different types of mixture models are reviewed. These are: linear, probabilistic, geometric-optical, stochastic geometric, and fuzzy models. A summary of the conception and formulation of each of these types of models is presented. A comparative analysis of the different attributes of the models is made. In a general sense, the linear, probabilistic, and fuzzy models are relatively simple while the geometric (geometric-optical and stochastic geometric) models are complicated, involving the incorporation of parameters of scene geometry. There is some difference in the number and nature of components that can be resolved with the different models. Available information is insufficient to categorize the models in terms of accuracy levels, but it is evident that mixture models produce more accurate land-cover estimation than conventional classification.