In the wind/photovoltaic hybrid power generation systems, with strong fluctuation of wind and photovoltaic resources, the stability of voltage becomes low. So voltage control is of the utmost importance for wind/photovoltaic systems. Voltage control is usually implemented by the optimization of local reactive power. The optimization of reactive power is a non-linear combination optimization problem with constraints and multi-extrema. It is very difficult to solve by the traditional methods. So an improved genetic algorithm (IGA) is designed for optimizing reactive power in the wind/photovoltaic hybrid power generation systems. Based on the simple genetic algorithm, the algorithm is improved with the method of coding, operators and termination conditions. Experimental results show that the improved algorithm can remarkably improve the convergent speed and calculational precision, and implement the reactive power and voltage control more efficiently.