Ensuring food security and human health is one of the main challenges of the 21st century for sustainable urban and rural development. Farmers use agricultural pesticides in large quantities annually to fight plant pests and diseases. One of the main problems in plant spraying is the drift of pesticide droplets which have a destructive environmental impact and pollute agricultural lands, animals, and humans. In this research, the volume median diameter (VMD) was chosen as a measure of the size of deposited particles in the non-target areas of spraying. This measure was obtained from image processing of water-sensitive papers that were placed in a wind tunnel to collect airborne and drifted particles. Different artificial neural network methods were developed to investigate the Factors affecting spray drift of pesticides. According to the results, the ANN model with topology 4-12-12-1, the root means square error (RMSE) 0.0388, and the coefficient of determination (R2) 0.9869 can forecast the volume median diameter more accurately. The sensitivity analysis for the neural network model showed that the influence of independent variables, including wind speed, spraying pressure, boom height, and nozzle type is 42%, 27%, 16%, and 15%, respectively. In the neural network model, the wind speed had the greatest effect on potential spray drift compared to other variables (42%). Based on results to control pesticide drift, the proper techniques of application of a sprayer by focusing on the use of spraying tools at lower wind speed and appropriate spraying pressure (Prevent high spraying Pressure), while limiting unintentional environmental damage (pesticide drift), increases agricultural productivity. This practical mode can lead to the sustainable development of agriculture and also ensure food safety and human health.
