Increasing energy consumption of manufacturing industry demands novel approaches to achieve energy conservation and emission reduction. Most of the previous research efforts in this area focused more on analyzing manufacturing energy consumption of a process or that of a machine tool with less concern on the system level of advanced machining workshop, especially a flexible manufacturing system. In this article, a new energy-saving approach of flexible manufacturing system is put forward based on energy evaluation model for integration of process planning and scheduling problem in flexible manufacturing system (flexible manufacturing system-integration of process planning and scheduling). First, complying with feature precedence and other technological requirements, flexible manufacturing system -integration of process planning and scheduling is mapped as an asymmetric traveling salesman problem of which operations are provinces and candidate operations are cities belonging to different provinces. To evaluate the performance of each solution of the asymmetric traveling salesman problem, energy consumption evaluation criteria for flexible manufacturing system-integration of process planning and scheduling are established and three energy efficiency indicators are also provided to perform further analysis on manufacturing energy consumption, that is, part energy efficiency, machine tool energy efficiency and feasible solution energy efficiency. Then, a mutation-combined ant colony optimization algorithm is proposed to solve the flexible manufacturing system-integration of process planning and scheduling which combined roulette and mutation selection methods to pick out the next candidate operation. The pheromone trails associated with edges are released by the so-far-best ant or the iteration-best ant probabilistically to both keep the search directed and avoid converging to the local best. Finally, a case study of flexible manufacturing system in advanced machining workshop is employed to demonstrate the feasibility and applicability of this approach in three different scenarios and compared with the "process planning then scheduling" approach; energy consumption obtained by the proposed method drops 10.7%.