The global optimization for complex functions method is in dire need of new computation. By combining the Powell method, which finds a partial optimum solution, with the uniform distribution design method, which uses evenly spaced test points at higher dimensional space, allows for a new algorithm which is capable of solving the global optimization function. This new algorithm is robust enough to search global optimal solution. Its time complexity is proportional to the number of initial points and to the square dimensions of the problem, and its spatial complexity is in proportion to the number of the initial points and the dimensions of the problem. Also, the algorithm has good parallel efficiency and will have broader potential applications.