Graph coloring is a NP-Complete problem. In this paper, based on the research result of Matrix genetic operator, inverse and dual combination operator, an optimization combination genetic algorithm is constructed through the inverse and dual combination genetic operator combined with matrix genetic operator to solve Graph Coloring problem. A transform coding between integer and binary is introduced to make good use of the combination operators. Fitness function based on constraint of Graph Coloring problem is designed, and the convergence of algorithm is proved. Better efficiency of the optimization combination genetic algorithm for solving Graph Coloring problem is verified compared with current genetic algorithm.