Prediction the dynamics of the level and structure of regional hunting resources is relevant from the standpoint of their rational use. Matrix models are widely used to make predictions on the dynamics of hunting animal populations. The algorithm of the modified P. H. Leslie matrix model with a correction matrix is used. The accuracy of predictions on the dynamics of hunting animal populations based on matrix models is improved by their verification. In the proposed study, model verification is considered not only as a method for determining the correspondence of the model to the corresponding modeling object, but also as a tool for clarifying model parameters under conditions of possible uncertainty of information. A retrospective verification of models of the dynamics of prediction the number of hunting animal populations is considered, the results of which are verified by prospective verification. On the basis of retrospective verification under the conditions of incompleteness of the available information, the parameters of the models are clarified. The proposed prediction algorithm works well with a steady increase in the population of hunting animals. In practice, there are regressive and unstable scenarios of the dynamics of the number of hunting animal populations. In a regressive and stable scenario, the proposed algorithm for predicting the number of hunting animal populations works well, but regressive results give predictions for a decrease and even degradation of the population. In this case, the prediction tasks change. For example, for determining the percentage of production of a given type of hunting animals while maintaining the population size or its insignificant growth. As a result of the research, predictions were made on the dynamics of the populations of the main hunting animals in the Yaroslavl region (moose, bear, fox, white hare, grouse and capercaillie) based on verified matrix models.