Abstract: Focused on a man-machine collaborative workshop in the customized manufacturing environment, the solution about system performance indexes is studied based on queuing network modeling. Considering multi-resource constraints in particular, an open queuing network model with finite buffer capacity is built to describe stochastic process of the system, and the state space and state transition balance equation of each node built by using the state space decomposition method. The steady-state probability of the system in each state is obtained by solving the state transition balance equation, and then the system performance indexes were approximately solved. A simulation model is built. Through the comparative analysis of the experimental results with the results of the queuing network model, the relative error of each system performance index is less than 11%, which verifies the accuracy and effectiveness of the solution. The influence of different parameters on the system performance is analyzed based on the queuing network model. It provides basis for further research on resources optimization of this type of workshop.