Aiming at high-dimensional multimodal optimization problems, traditional evolutionary algorithms have shortcomings, such as low convergence speed and solution precision. A global optimization algorithm based on Memetic algorithm using global search strategy and local search strategy is proposed to resolve the high-dimensional problem. The global search strategy is a multi-model parallel differential evolution. An improved Simulate Anneal Arithmetic is used for local search strategy. The improved Memetic algorithm inherits advantages of the differential evolution algorithm to discover the global optimal solution and overcomes the deficiencies of the differential evolution algorithm. Finally, four benchmark functions are used to test this algorithm. Experimental result illustrates that it has some advantages in convergence velocity, solution precision, and stabilization.