This paper deals with the maintenance optimization of a system subject to a stressful environment. The system deterioration behaviour can be modified by the environment; the degradation mode can change due to the random evolution of the stressful environment. Reciprocally, the environment conditions can be influenced by the system state and as a consequence, a change in the environment can be an indicator of the system state. This paper describes a condition-based maintenance decision framework to tackle the potential variations in the system deterioration, and especially in the deterioration rate, and the new information on the system state given by the evolution of the environmental variables.