The invention proposes an improved genetic algorithm for flexible workshop scheduling, and the algorithm relates to the technical field of workshop scheduling, and specifically relates to the technical field of flexible workshop scheduling. The invention aims at the problems that a conventional genetic algorithm is complex in coding mode, is difficult for decoding, is weaker in search and development capability and is liable to be mature early and a non-feasible solution is liable to appear in the operation of a genetic operator. Compared with a conventional algorithm, the algorithm has the following improvements that 1, the coding is just performed on one chromosome, a coding chromosome gene consists of a ternary array (i, j, k), the coding mode is simple and convenient, and there is no need of decoding; 2, a positioning method is employed for selecting equipment for the process according to two different rules, and three known effective scheduling rule is employed for process arrangement; 3, crossing and mutation operations employ a genetic operator based on process priority protection; 4, before mutation, the probability of individual and genetic mutation is calculated through a formula, thereby achieving more accordance with the natural law. The algorithm is high in practicality.