Abstract: Task scheduling in heterogeneous environment is a NP problem, it is concerned about the large-scale resource and task scheduling, requires the scheduling algorithm is highly efficient. As the number of tasks and resources to increase the number of genetic algorithm to show the shortcomings of slow convergence. In order to overcome its shortcomings, this paper improved genetic algorithm based on the increased packet processing and load balancing strategy, a hybrid genetic scheduling policy, simulation experiments show that the hybrid genetic algorithm-based scheduling policy than the traditional scheduling strategy better performance, the algorithm more complex, heterogeneous environment, the system can better meet the characteristics of time and resources to minimize the overhead issue.