Aiming at the shortcomings of the golden jackal optimization algorithm (GJO) in solving complex optimization problems, such as slow convergence speed, poor optimization accuracy, and easy falling into local extremes, a mixed-strategy improved golden jackal optimization algorithm (MSIGJO) was proposed. Firstly, a Tent mapping fused with inverse learning initialization strategy was introduced to enhance the diversity of the golden jackal population and improve the quality of the initial solution of the algorithm. Secondly, a dynamic adaptive energy factor was used to coordinate the ability of the algorithm to search globally and develop locally in-depth, and once again, a sinusoidal cosine algorithm was fused with a helical search strategy guideline mechanism in the exploration phase to improve the convergence speed of the algorithm. Finally, the algorithm was enhanced with backward learning and differential variation strategies to avoid the algorithm falling into local optimization. Using the improved algorithm proposed the basic optimization algorithm, and the existing improved golden jackal algorithm for simulation experiments on different test functions, through the simulation results of the comparative analysis can be obtained that MSIGJO convergence, robustness, Wilcoxon rank-sum test and time complexity are better than other algorithms, with good convergence speed, solution accuracy, and robustness. MSIGJO was used to solve engineering optimization problems, which verifies the effectiveness and feasibility of MSIGJO algorithm in solving real engineering problems.