Xinjiang is one of the most important highquality cotton production areas in China, and sprayers with horizontal boom and hang boom are widely used in ripening and defoliation of the cotton. Due to the close planting cultivation of cotton, in the lower part of the cotton canopy, the spraying coverage of the defoliant is low and defoliation effect is poor, which seriously affects the cotton quality. In order to improve the spraying coverage rate of the defoliant in the middle and lower cotton, a scheme for 〖JP3〗optimizing the parameters of the divider was put forward, and it was designed and developed by using Box-〖JP〗Behnken. The parameters, such as top rake of the crop divider, ground clearance, field speeds were taken as the influencing factors, and spray coverage and resistance of crop divider were used as test indices in the experimental study, obtaining the test indices though field test with test equipment designed. By using the ellipsoidal basis function neural network (EBFNN) the relationship between the indices and test factors was approached, then accurate and reliable approximation model was established. Then the multiobjective genetic algorithm was used to optimize the coverage rate and resistance of crop divider based on this approximate model in Isight software platform, the optimal parameters combination was obtained through determining the weight coefficient of the optimized solution set. Best combination of test parameters were listed below: ground clearance of the crop divider was 210mm, the top rake of the crop divider was 12°, and the operation speed of the sprayer was 4km/h. Field experiments were carried out under this condition, the results show that the average droplet coverage on cotton canopy was 22.49%, compared with the model prediction, the error range was less than 10.89%, and the root mean square value of grain resistance test was 70.9N, the error range was less than 7.78%. It can provide a reference for cotton crop divider design and spraying parameters optimization of boom sprayer, and also greatly promote the progress of cotton defoliation harvesting mechanization.