Job Shop scheduling problem, the essence of which is the resources scheduling problem, which has been proved to be a complete NP-hard problem. It has importantly realistic effect on further research, and has become a hot spot of research now. According to the practical Job Shop, as equipment resources are not unique, there are several machine tools with high frequency, while the number of that of low frequency is only one; the working procedure of processing components are also quite different, so we have put forward the Genetic Algorithm considering the sequence and, simultaneously, the machine choice. For reaching the shortest producing period, this method adopts Gemini string to encode, combining with the characteristics of the resources scheduling problem, and designs the unique way of Crossover and Mutation, meanwhile, it shows that the algorithm is effective through a specific example simulation analysis.