With the continuous development of social economy, enterprises have higher and higher requirements for production efficiency and product quality. How to complete the task assignment under the basic constraints of ensuring production capacity and low cost is the main problem faced by the manufacturing manufacturing industry. This paper presents a fault-tolerant strategy for distributed systems. Firstly, the paper gives the research background and significance. Secondly, it introduces the layered scheduling model based on EDF and the distributed system with weight weighting designed by particle swarm optimization algorithm. The simulation scheme is tested and analyzed by MATLAB software to verify the results. Finally, it is concluded that the scheme can fully ensure the stability of the scheduling process in the actual working environment. The stall error rate is also guaranteed to be within the range of failure, so it can be widely used in various fields.