In order to improve the quality and efficiency of coffee blend, a method based on interactive genetic algorithm (IGA) for it was studied. The blending solutions were used as individuals, with the improved adaptive crossover and mutation probability, and the interactive interface provides 10 indexes for individual fitness, and then five users participated the optimization. The results showed that the method was easy to operate, a consistent satisfactory solution could be obtained within 4～5 generations, with the difference of component ratio less than 5% and the fitness more than 85. Compared with the traditional cup testing, the results showed that the IGA optimization method was superior to the traditional one. Based on the successful application in evolutionary optimization of IGA, this research provided a new intelligent way for coffee blends.