Abstract: Aiming at solving the problems of trapping into local extremes easily in the process of optimization, there is a fruit fly optimization algorithm based on combinatorial variation proposed.In this paper the combined mutation strategy is introduced when the algorithm falls into the local extreme.In the basis of this, the flight direction is updated and the population escape ability is enhanced, Thus, the algorithm continues to iterate to find the global extreme value.The experimental results show that the fruit fly optimization algorithm based on combinatorial variation is a good way to balance the algorithm's local search ability and global mining ability, and the algorithm is improved in terms of convergence speed and convergence accuracy.