Proper inventory planning should incorporate factors changing over time, since static factors are not robust to this apparent variability. In models of inventories is necessary to recognize the great demand uncertainty. This paper reviews the state of the art of the most significant developments related to inventory models, especially those who consider dynamic demands in time. In addition, demand forecasting models and some techniques for optimizing inventories are analyzed, considering costs and service levels, among others. In the literature review, gaps have been identified related to the treatment of multivariate inventories as well as the use of Bayesian statistics for the purpose of optimization and the development of demand forecasts.