Abstract: Aiming at the optimization of branch air volume in mine ventilation network, taking the minimum total power of mine ventilation network as the goal.Based on the balance equation of stroke volume, wind pressure balance equation, branch resistance equation and fan characteristic curve equation, a multi group adaptive particle swarm optimization algorithm (MA-PSO) is proposed to optimize the mine ventilation network.Firstly, the random population is generated and pretreated, and the adaptive value is sorted from high to low. Secondly，taking local optimal solution as the center, the local optimal solution and the average value of the Euclidean distance of other particles as the radius, and the population is divided into five subpopulations, and then the topology item is introduced into the speed updating formula. With the population exchange factor, the particle is searched in the solution space in the subpopulation, and the diversity of the population is ensured, thus the speed of evolution and convergence is accelerated. Finally, the adaptive weight and redundant particles are used to initialize the elimination strategy to improve the searching ability and learning ability of the algorithm. The simulation results show that the algorithm has better multi-modal optimization rate, faster convergence speed and higher convergence precision. The total power consumed by the optimized ventilation system is 26.78% lower than that of the previous, and the energy saving effect is remarkable.