Traditional control theory does not consider some uncertain factors in the industrial process,which has a significant impact on energy loss and system accuracy in production process.This paper considers the data-based optimization control strategies for non-Gaussian stochastic systems.Kernel density estimation was used to estimate the output probability density functions (PDFs) on the basis of collected output data.Firstly,the performance index function was established based on the control objectives.Secondly,the performance index function was optimized by a genetic algorithm.The simulation takes the grinding system as a model and uses PDFs to characterize the particle size distribution.Simulation results show that the genetic algorithm-based stochastic distribution control for non-Gaussian systems can effectively achieve the control target of the stochastically distributed control system and provides reference for practical industrial production.