Abstract: The prediction of crop development stage plays a critical role in agrometeorological service. The simulation results of four crop development models were compared, which can provide significant suggestions for crop development prediction in Northeast China. Based on the meteorological data and the observational data of development stages for maize, rice and soybean, parameters of four models were calculated and validated by both internal and external data using a simulated annealing method. The results show that in the course of parameter localization, the simulating effects of Gao’s Model and Shen’s Model are better, and root mean square errors (RMSE) are 3.31 days and 3.72 days, respectively. In the course of model validation, Shen’s Model has a better simulating effect and average of RMSE is 5.22 days.

Key words: Rice, Maize, Soybean, A simulated annealing method, Development stage, Model