Abstract: Probability and uncertainty of process plans are important issues for intelligent CAPP systems. The model described in the paper can balance the scale and efficiency of intelligent planners, and build a man machine cooperation environment. Procedure termination, planning navigation, task generation ＆ reorganization, data dispatch, and planning state interpretation make up the internal transactions of a CAPP system. Their combination is characterized by varied forms, such as recursion, loop, sequence, concurrence, and backtracking.The knowledge about feature, part, process resource,and enterprise is used to plan a manufacturing process. The model improves the system’s reliability and operation efficiency, and its key technique is implemented in a GS CAPP system.