To improve the performance of optimization algorithms in complex function optimization scenarios, this paper proposes an improved Dhole Optimization Algorithm (FETDOA). Based on the original Dhole Optimization Algorithm (DOA), FETDOA adopts Fuch chaotic mapping to enhance the uniformity of the initial population and expand the search coverage. The tangent flight strategy introduces controllable perturbations in the orthogonal subspace of the prey attraction direction to ensure convergence and strengthen local exploration, thus alleviating premature convergence and local stagnation in complex multimodal problems. The improved Experience Exchange Strategy (EES) enhances information sharing among individuals and improves the efficiency of collaborative search, ultimately achieving fast and high-precision convergence. To verify its performance, FETDOA is compared with eight state-of-the-art algorithms, and the Wilcoxon rank-sum test is used for statistical significance analysis. Experimental results show that FETDOA is significantly superior to the comparison algorithms in terms of mean and standard deviation on most test functions. Finally, three engineering cases are adopted to demonstrate the excellent performance of FETDOA in engineering applications.In summary, FETDOA has stronger convergence accuracy and stability in complex function optimization, providing an efficient solution for practical problems such as engineering optimization and parameter tuning.