This paper deals with a continuous-time stochastic control model designed for planning production and preventive maintenance in a flexible manufacturing system. This model extends previous formulations due to Olsder and Suri or Kimemia and Gershwin by including an age dependent machine failure rate and by allowing the control to influence some jump rates (viz. the preventive maintenance activities). By using an adaptation to the case of piecewise deterministic systems, of the approximation technique initially proposed by Kushner in the realm of the optimal control of diffusions, one shows how it is possible to compute the optimal control for a two-machine system.