It is known that territories of nature plague foci are non-equal by the degree of epizootic situation tension. There are locations (sectors of primary regions) where plague has not been registered or has been registered rarely during all the time of surveillance. In this connection, it is necessary to reestablish the borders of plague focus territory for excluding the sectors without plague manifestations or the areas where plague occurs periodically. However, mathematically reliable criteria to substantiate such decisions are currently unknown to the authors. Objective was to propose evidence-based criteria for the exclusion of certain areas from plague focus territory using probabilistic approach. Material and methods. Surveillance data on Central-Caucasian natural plague focus for the period of 28 years (1987-2013) were used. For each sector of primary zoning in the focus a multiplicity of surveillances and multiplicity of plague epizooty reports was calculated. Some nonparametric techniques (Spearman correlation, quantile analysis) and concepts of probability theory were used for statistical processing. Results and discussion. By the example of Central-Caucasian high-mountain natural plague focus, the method for probability calculation for manifestations of epizooties in the sectors of plague focus is put forward. The evidence-based criteria for the exclusion of certain areas from plague focus territory using probabilistic approach are proposed. Algorithm for excluding the sectors of primary zoning from the surveillance scheme taking into account different levels of significance is presented. The ranking of the focus by the probability of epizooty detection in a sector and probability of surveillance can be utilized alongside the index of frequency of epizooties when planning epizootiological survey of the focal territory.