Abstract: Air-assisted boom spraying has proven to have a positive effect on drift reduction. However, its effects with certain operational parameters of air-assisted boom sprayers are varied depending on natural wind speeds, outlet velocities of air curtain, spraying flow, spraying angle and etc, thus the spray drift rate and deposition rate are not stable in the field working. For example, if wind speed is low, and airflow and outlet speed of wind duct are high, which not only increases the power consumption of the fan, but also directly blew the droplets to ground and formed serious spray loss. On the contrary, when the wind speed is high, the speed of air-assisted flow is not enough to overcome the influence of wind, the spray droplets is drifted. To provide the control parameters to achieve precise control model with pesticide effect, this paper, leveraged by a three-dimensional multiphase flow computational fluid dynamics (CFD) model with the consideration of wind speeds, inlet speed of air curtain, spraying flow and spraying angle, simulates the coupling interaction of natural wind (continuous phase), the air curtain (continuous phase) and the droplets (discrete phase) of the air-assisted boom sprayer to study the droplet drift characteristics. The uniform experimental design took into the four factors with five levels of L2- deviation of the uniform design table U25 being applied to arrange the simulation scheme. Two criteria, the drift rate and the downward velocity under the duct 0.5m, were used to evaluate the spraying performance under the computing utilizing CFD simulation results. The simulation results were collected as training samples, and the multivariate relevance vector machine (MRVM) regression method was utilized to establish the 4-inputs-2-outputs spraying drift model accounting for the varying natural wind. The CFD simulation and the MRVM model only considered the four factors which influenced the spraying effect, however the vertical distance when the dense degree, crop of nozzle and the crop canopy were not at the same time, influence of spray effect was also affected by these parameters. Therefore, a fuzzy inference system model considersing crop canopy dense degree and vertical distance between nozzles and canopy was established to correct the 3 control parameters. According to the experiment, spray system analysis and expert experience, 11 fuzzy rules with Gauss membership function were set up. By using the fuzzy logic toolbox, fuzzy inference system was defined to obtain the mapping between input and output. In order to quantitative analysis the modeling quality of the droplet drift characteristics, the 3MQ - 600 type air-assisted boom sprayer was used in the droplet drift test. Model tests showed that the mean absolute percentage error of the drift rate was 2.56%, and the spraying drift test of air-assisted boom sprayer prototype with natural wind disturbance had validated that the measured and predicted drift loss rate average error of 8.92%, which still showed the same interaction rule with the built spraying drift predictive model. This study provided the active control system with spraying ant-drift and droplet deposition effects-oriented control and decision-making model.