Abstract: The objective function of back analysis of displacement parameters of the elasto-plastic model for rock mass has many minimums. When the conventional mathematical optimization methods are employed to solve the problem, the search often falls in one of the minimums and cannot go any further. In order to avoid the situation, the principles of genetic algorithm have been utilized to establish a genetic-algorithm-based approach to the back analysis problem. The approach has been used to inverse the mechanical parameters of a prescribed elastoplastic model and the accuracy of the results has reached ninety percent.