Abstract: Orthogonal experimental was carried out using the tomato dry weight as experimental objective, and the impact of environmental parameters (temperature, relative humidity, light intensity) on tomato dry weight was analyzed through experiment. A neural network calculation model was established based on experimental data and the test were made by using MATLAB software, and to better verify effectiveness of the approach, a 10-fold cross validation method was used. Through 10-fold cross validation, model achieved the predictive accuracy of 95.32%. The results showed that the predicted values were in good agreement with the experimental values. This method has high prediction precision, which can be used theory instruction in environment control, and the predicted dry weight of tomato with the BP neural network method was feasible．

Key words: tomato, dry weight, BP neural network, prediction