Abstract: By establishing a multi-objective canal system optimal water distribution model, the canal system water distribution process was optimised to provide a scientific basis for water-saving irrigation in Xinjiang agriculture. Taking the Xindong main canal in the Wujiaqu irrigation district of Xinjiang Province as the research object, the multi-objective particle swarm optimisation algorithm was used to establish a multi-objective canal system optimization water distribution model with the optimisation objectives of minimum leakage loss and stable water flow in the distribution channel, and compared with the backtracking search algorithm with regard to the water distribution time and the amount of leakage loss. It was shown that the multi-objective canal optimisation distribution model was able to reduce the distribution time from 15 days to about 11 days with a canal distribution of 0.715 8 million m~3, which was a reduction in the distribution time from 14 days to 11 days and a reduction in seepage of 0.083 9 million m~3 when compared to the backtracking search algorithm. The multi-objective particle swarm optimisation algorithm was used to optimise the water distribution process of the Xindong main canal in Wujiaqu irrigation district of Xinjiang, which not only satisfied the optimisation objective, reduced canal leakage and shortened the distribution time, but also was more practical than the backtracking algorithm.