Abstract:On the basis of the principle which the function of high nonlinearity is not appropriate to linearize,the limitation of traditional methods in solving parameters of nonlinear regression estimation is analyzed. A new method based on genetic algorithm is proposed, which adopts uniform design means to combine genetic operator, conduct numerical experiments and then obtain optimal estimation of parameters of regression estimation. The results show that there is no need of good initial values of parameters when genetic algorithm is adopted, both the capability of global search and the robustness of genetic algorithms exceed that of traditional algorithms.