This paper presents a hybrid self-adaptive orthogonal genetic algorithm (HSOGA) based on orthogonal experimental design method for solving global optimization problems. In HSOGA, the orthogonal experimental design method is utilized to design crossover operator, and as a result, a self-adaptive orthogonal crossover operator is proposed. The self-adaptive orthogonal crossover operator self-adaptively adjusts the number of orthogonal array’s factors and the location for dividing the parents into several sub-vectors according to the similarity of the two parents, in order to produce a small but representative set of points as the potential offspring. In addition, in HSOGA the self-adaptive orthogonal crossover operator is also adopted to generate an initial population that is scattered uniformly over the feasible solution space in order to maintain the diversity. Moreover, a local search scheme is incorporated into HSOGA in the purpose of enhancing the local search ability and speeding up the convergence of HSOGA. HSOGA is tested with fourteen benchmark functions. The experimental results suggest that HSOGA is generic and effective.