In order to improve the efficiency of dense assembly process, this paper establishes a mathematical model for its assembly process, obtains the multi-objective path planning problem in the assembly process, improves the ant colony algorithm according to the roulette principle, overcomes the problem that the algorithm is easy to fall into the local optimal solution, optimizes the assembly trajectory, explores the optimal path of dense assembly process, and improves the efficiency of assembly process.