Abstract: In order to obtain the impact of different monitoring and sampling intervals on crop model performance, and avoid the subjectivity and randomness in the existing research and application, the transpiration rate and CO2 exchange rate of tomato leaf were modeled by two common methods with three kinds of sampling intervals, then the prediction error of these models were compared. The results showed that the prediction errors of transpiration rate model and CO2 exchange rate model were different respectively with the sampling intervals of 15, 30 and 60 minutes, the prediction ability of the GA-BP neural network model was generally superior to the pure quadratic regression model, but the conclusions were consistent, showing that the most appropriate sampling interval was 30 minutes for transpiration rate model and 15 minutes for CO2 exchange rate model respectively. The results of this study provided the basis for the setting of monitoring and sampling intervals, the selection of experimental data and the application of the modeling methods, which was of great significance for the research and application of crop model.