Antlion optimization (ALO) is an efficient metaheuristic paradigm that imitates antlion’s foraging behavior when they search for the ants. However, the conventional variant appears to encounter difficulties in avoiding local optima stagnation and slow convergence speed in dealing with complex problems. Hence, there are problems in the performance that need to be mitigated. To alleviate these shortcomings, an improved variant called Lévy orthogonal learning ALO is developed, which enhances the efficacy of the core method with orthogonal learning strategy, Levy flight, and primary core mechanisms. To measure the effectiveness of the new method, it is compared with the basic version, variant called Levy flight ALO, and variant called orthogonal learning ALO using thirty benchmark functions from IEEE CEC 2017. Also, it is compared with 15 well-known metaheuristic algorithms. Empirical results have shown the superiority of the proposed algorithm in solving the majority of test functions in terms of solution quality and convergence speed. To further validate the efficacy of the enhanced algorithm, it is applied to common practical engineering problems with constrained and unknown search spaces. The obtained results vividly demonstrate that the proposed algorithm provides satisfactory results for solving these problems.