Image thresholding has received widespread attention in recent years, and corresponding intelligent optimization algorithms have been used to improve the search efficiency of its optimal threshold. This article proposes a hybrid multi strategy improved Wild Horse Optimizer (IWHO), which verifies its performance in optimal threshold search and analyzes its performance in multilevel image segmentation. During the initialization phase, IWHO uses Tent chaotic mapping to initialize the population. During the population search process, linear weight methods are introduced to enhance its local optimal search ability. Finally, the global search performance of IWHO is improved through the Cauchy mutation factor. In order to improve the practical application value of the IWHO, its effectiveness in medical image segmentation was tested and compared with the Wild Horse Optimizer (WHO), improved Satin Bowerbird optimization algorithm (ISBO), Coyote optimization algorithm (COA), Ant Colony Optimization based on reverse learning (MALO), and improved Whale Optimization Algorithm (MWOA). The experimental results indicate that IWHO has shown excellent performance in image segmentation quality and comparison of experimental data from similar methods.