Mill/turn machining center can perform several operations simultaneously with multiple spindles and power turrets, which means worksteps sequencing should be integrated with scheduling. According to the characteristics of mill/turn machining center, an improved genetic algorithm to optimize worksteps sequence was proposed. A 0-1 mixed integer programming model was set up to support the algorithm with object function and constraints. In order to present so many parameters of the 0-1 mixed integer programming model, coding strategy was developed based on data structure instead of integer or symbol. According to the similarity of genes, two novel genetic operators named multi-probability crossover operator and multi-probability mutation operator were also introduced. Consequently, these genetic operators insure that chromosomes consisted of genes with more similarity could transmit their characteristics to next generation with higher probability. The effectiveness of the proposed algorithm was verified by a real case using dual spindles and twin power turrets mill/turn machining center. The results also indicated that the improved genetic algorithm gained a better performance than the conventional genetic algorithm.