This paper develops a performance evaluation procedure for a multi-line production system with different station failure rates. Considering reworking actions, we propose a graphical-based methodology to model the production system as a stochastic-flow network. First, a transformation technique is utilized to build the production system as a production network. Second, a simple algorithm integrating decomposition technique is proposed to generate the minimal capacity vectors that stations should provide to satisfy the given demand. We evaluate the probability that the production network can meet the demand in terms of the minimal capacity vectors, where the probability is referred to as the system reliability. A tile production system is utilized to demonstrate the performance evaluation procedure. A further decision making issue is discussed based on the derived system reliability.