It was shewn by applying different models ( Logistic, Gompertz and Weibull) to data obtained by field disease investigation that Weibull model described the progress curves of Chinese-Small-Cabbage Mosaic ( mainly caused by Turnip Mosaic Virus and Cucumber Mosaic Virus ) the best, and the Logistic better than Gompertz model. Nonlinear method, Gauss-Newton algorithm and Marquardt' s algorithm, was proposed for the fitting of two-parameter equations of plant disease progression. A comparison was made of nonlinear and linear method by using 94 sets of plant disease progress data o f 18 disease systeme; results showed that the nonlinear method was better than the linear method in all casas.