This paper presents a joint optimal inventory control and preventive maintenance strategy for a randomly failing production unit which supplies an assembly line operating according to a just-in-time configuration. The production unit is submitted to a maintenance action as soon as it reaches a certain age T or at failure; whichever occurs first. In order to palliate perturbations caused by breakdowns and by planned maintenance actions both with random durations, a buffer stock is built up to ensure the supply of the assembly line at a constant rate. A mathematical model is developed to evaluate the average cost per time unit of the proposed strategy. Due to the complexity of the mathematical model and to some approximation assumptions, a second approach based on simulation and experimental design is developed. It allows us to simulate the proposed policy and to generate a second order response surface for easy computation of the optimal values of the decision variables which are the age T at which preventive maintenance must be performed, and the buffer stock level h.