<p>Floating offshore wind (FOW) turbines have emerged as a potential means to harness wind energy in deep ocean waters where fixed-base turbines are infeasible. However, replacement of major turbine components is challenging due to harsh environmental conditions and limited availability of specialized resources, such as jack-up vessels (JUVs). Unlike fixed-base turbines, FOW turbines can be towed either to shallow waters or to onshore facilities for replacement activities. In this paper, we present a stochastic optimization model to minimize the expected total, discounted cost of replacing fixed-base and FOW turbines when the turbines are subject to randomly varying environmental conditions. The shallow and deep water environment states are modeled as a bivariate Markov chain which exhibits (weak) positive correlation at each time step. A Markov decision process (MDP) model is created to jointly optimize the timing of replacement and JUV requesting decisions over an infinite planning horizon. The model incorporates stochastic dependence arising from degradation processes influenced by correlated environment states, as well as economic dependence induced by sharing of a JUV across turbine types. We characterize the structure of optimal policies and show that both replacement and JUV requesting decisions exhibit threshold-type behavior with respect to turbine degradation. Finally, we provide sufficient conditions under which these policies are monotone in the environment state.</p>