Focusing on the issue that the novel metaheuristic optimization algorithm-Gray Wolf Optimizer (GWO) is easy to fall into local optimum when it is searching for the global optimal solution, thereby its ability was enhanced to obtain the global optimal solution. The fundamental principles and modeling processes of GWO were introduced firstly. On this basis, combined with the advantages of self-organized criticality theory, the Improved Extremes Optimization (IEO) algorithm was proposed. Then the IEO was integrated into the GWO model to construct the Self-Organized Critical (SOC) optimization algorithm named IEO-GWO. By adopting 23 benchmark test functions to implement a comprehensive comparison with traditional optimization algorithms in optimization performance, the superior ability of IEO-GWO model in searching global optimal values was verified.