An enhanced dung beetle optimization algorithm (EDBO) is proposed for solving nonlinear optimization problems with multiple constraints in manufacturing. The algorithm improves the rolling phase by removing the interference of the worst value and coupling the current solution with the optimal solution, while retaining the original formulation's advantages. To address the issue of the dancing phase focusing only on current solution information, the global optimal solution is introduced to guide the dung beetle, and a stochastic factor is added to the optimal solution. The foraging phase introduces the Jacobi curve to enhance the algorithm's ability to escape local optima and avoid premature convergence. The performance of EDBO is tested using the CEC2017 function set, and statistical tests (Wilcoxon rank-sum and Friedman tests) confirm its effectiveness. Experimental results show that EDBO has strong optimization accuracy and stability. It is applied to four engineering optimization problems, demonstrating good adaptability and robustness. The EDBO algorithm outperforms other algorithms in convergence speed and solution quality across various dimensions and functions. The algorithm's improvements enhance its global search capability and local exploitation, making it suitable for complex engineering optimization tasks.An enhanced dung beetle optimization algorithm (EDBO) is proposed for solving nonlinear optimization problems with multiple constraints in manufacturing. The algorithm improves the rolling phase by removing the interference of the worst value and coupling the current solution with the optimal solution, while retaining the original formulation's advantages. To address the issue of the dancing phase focusing only on current solution information, the global optimal solution is introduced to guide the dung beetle, and a stochastic factor is added to the optimal solution. The foraging phase introduces the Jacobi curve to enhance the algorithm's ability to escape local optima and avoid premature convergence. The performance of EDBO is tested using the CEC2017 function set, and statistical tests (Wilcoxon rank-sum and Friedman tests) confirm its effectiveness. Experimental results show that EDBO has strong optimization accuracy and stability. It is applied to four engineering optimization problems, demonstrating good adaptability and robustness. The EDBO algorithm outperforms other algorithms in convergence speed and solution quality across various dimensions and functions. The algorithm's improvements enhance its global search capability and local exploitation, making it suitable for complex engineering optimization tasks.