Abstract: An optimal schedule of constrained resource on the critical chain can effectively improve the operational efficiency of the system. This is one of the important research issues in critical chain project management. Existing research on constrained resource mainly focuses on updating of resources rather than on the substitutability of resources. Therefore, the scheduling of partially substitutability constrained resource is studied to improve resource utilization and cut down the expense etc. Firstly, the scheduling problem is described as the parallel allometric machines with machine eligibility restriction by a triplet. With the goal of minimizing tardiness, a mathematical model is established. Then, a hybrid rescheduling strategy is used with improved particle swarm optimization to build an algorithm. Finally, simulated examples are designed, by comparing with particle swarm optimization (PSO) and GA to demonstrate the feasibility and superiority of the solution.