This paper proposes an improved Humboldt squid optimization algorithm (GMHSOA) based on genetic mutation to address the issues of multiple adaptive parameters and high time cost that the generation of egg positions during the exploitation of the newly proposed Humboldt squid optimization algorithm (HSOA). The proposal of this algorithm starts from the essence of the optimization algorithm exploitation stage, that is, the generation of egg group positions requires inheriting the optimization results of parents on the one hand, and effectively avoiding the problem of algorithm falling into local optima on the other hand. In other words, the egg group also needs to undergo mutation or mutation to generate new positions and search for the optimal solution again, which precisely reflects two aspects of genetics and mutation. In order to verify the performance of the proposed algorithm, the GMHSOA algorithm was compared with HSOA, PSO, COA, PDO, SSA, and BES algorithms based on 23 benchmark test functions in terms of algorithm convergence and time cost, and good results were obtained. Therefore, the algorithm proposed in this article has certain advantages.