Abstract: A new approach for back analysis of mechanical parameters of dams combining BP neural network with uniform design and genetic algorithm was established.Firstly,basic genetic algorithm was improved for very good global and local searching capability.The genetic neural network using the improved genetic algorithm as its learning algorithm was established to overcome the shortcoming of BP algorithm.Secondly,the sample of material parameters was designed by the uniform design method,and the sample of the calculated displacement of dams was obtained by use of the finite element method.Through these samples,the above genetic neural network was trained to describe the sophisticated nonlinear relationship between displacement and material parameters of dams.Finally,the actual dam displacement was input into the trained genetic neural network to obtain the real material parameters.As an example,the elastic moduli and the linear dilatable coefficient of concrete of the dam body and the elastic moduli of rockmass under the base of Geheyan Dam were back-analyzed by the above method.It was shown that the method could shorten the time of back analysis and improve the efficiency and accuracy of back analysis.