The paper introduces a multi-strategy improved dung beetle optimization algorithm (MDBO) to address the limitations of the original DBO algorithm. The original DBO algorithm, inspired by the behaviors of dung beetles, has been shown to have issues such as low population diversity, susceptibility to local optima, and slow convergence when solving complex optimization problems. To overcome these challenges, the MDBO algorithm incorporates several improvements: (1) Latin hypercube sampling is used for better population initialization, (2) a novel differential variation strategy called "Mean Differential Variation" is introduced to enhance the algorithm's ability to escape local optima, and (3) a strategy combining lens imaging reverse learning and dimension-by-dimension optimization is applied to the current optimal solution to improve its quality. The MDBO algorithm is tested on standard benchmark functions from CEC2017 and CEC2020, demonstrating superior performance in terms of optimization accuracy, stability, and convergence speed compared to other classical metaheuristic algorithms. Additionally, the algorithm is successfully applied to three real-world engineering problems: extension/compression spring design, reducer design, and welded beam design, validating its effectiveness in solving complex engineering optimization challenges. The MDBO algorithm is also analyzed in terms of computational complexity, showing that it has the same complexity as the original DBO algorithm. The experimental results show that MDBO outperforms other algorithms in most test functions, with significant differences in several cases. The algorithm's performance is further validated using the Wilcoxon rank-sum test, which confirms the significance of the differences between MDBO and other algorithms. Overall, the MDBO algorithm demonstrates superior optimization capabilities, enhanced stability, and faster convergence speed, making it a promising solution for complex optimization problems.The paper introduces a multi-strategy improved dung beetle optimization algorithm (MDBO) to address the limitations of the original DBO algorithm. The original DBO algorithm, inspired by the behaviors of dung beetles, has been shown to have issues such as low population diversity, susceptibility to local optima, and slow convergence when solving complex optimization problems. To overcome these challenges, the MDBO algorithm incorporates several improvements: (1) Latin hypercube sampling is used for better population initialization, (2) a novel differential variation strategy called "Mean Differential Variation" is introduced to enhance the algorithm's ability to escape local optima, and (3) a strategy combining lens imaging reverse learning and dimension-by-dimension optimization is applied to the current optimal solution to improve its quality. The MDBO algorithm is tested on standard benchmark functions from CEC2017 and CEC2020, demonstrating superior performance in terms of optimization accuracy, stability, and convergence speed compared to other classical metaheuristic algorithms. Additionally, the algorithm is successfully applied to three real-world engineering problems: extension/compression spring design, reducer design, and welded beam design, validating its effectiveness in solving complex engineering optimization challenges. The MDBO algorithm is also analyzed in terms of computational complexity, showing that it has the same complexity as the original DBO algorithm. The experimental results show that MDBO outperforms other algorithms in most test functions, with significant differences in several cases. The algorithm's performance is further validated using the Wilcoxon rank-sum test, which confirms the significance of the differences between MDBO and other algorithms. Overall, the MDBO algorithm demonstrates superior optimization capabilities, enhanced stability, and faster convergence speed, making it a promising solution for complex optimization problems.