According to the important status and practical values of geographical factors and genetic factors in protein content estimation of winter wheat, a method analyzing the relationship between factors and protein content was proposed in this paper. In the method, genetic, geographical factors were chosen as independent variables, multivariate nonlinear regression model was introduced to estimate protein content. Firstly, construct one-dimensional nonlinear regression model based on each factor. Secondly, analyze and select sensitive factors on basis of the above simulations. Finally, couple the selected sensitive factors into multivariate nonlinear regression model for protein content estimation. Winter wheat was chosen as experimental subject. The results showed that, at province scale and county scale, the proposed method could not only effectively improve simulation efficiency, but also enhance whole estimation accuracy.