We study the integration of order acceptance and capacity planning in multi-project environments with dynamically arriving projects. We model this planning problem as a continuous-time Markov decision process to determine long-term optimal decisions. We examine whether macro-process planning should be performed before or after order acceptance. We characterise the structure of optimal policies, and explore the dependence on a number of parameters such as project payoff, project cost and order arrival time. We also look into the effects of set-up costs and the use of non-regular capacity.