The invention relates to a risk evaluation method of an electric power system network, and belongs to the technical field of data processing. The method comprises the following steps: (1) indicating a total loading value inside an area by using a concentrated load a, and defining all distributed power supplies in the area as a random variable set A in the area; (2) dividing the boundary conditions of the area according to the load in the area and correlation among the distributed power supplies; (3) checking the power system network according to the boundary conditions determined in the step (2), and determining the structure of the power system network; (4) obtaining the history time series data of each variable of six boundary conditions in the step (2), and forming data pairs using serial numbers and power numerical values as labels according to time serial numbering; and (5) calculating power system tide equations under the six boundary conditions in the step (2) respectively by utilizing a Monte Carlo method, and evaluating the risks of the power system network. The method provides effective bases for power network evaluation planning, and much unnecessary work is reduced.