In the present paper, we consider the demand management decisions of a manufacturer facing stochastic demand. While in the previous literature, either the order release decisions are fixed upon arrival or a single-stage production environment is assumed, we make online order acceptance and order release decisions in a multi-stage production system. After describing the problem formally as a stochastic dynamic program, we develop a bid-pricebased revenue management approach in which the order acceptance and order release decisions are made based on previously computed bid prices and feasibility according to a linear program describing the multi-stage production system. A numerical study shows the good performance of the approach compared to an ex-post optimal solution in various supply and demand settings as well as the benefits relative to existing models, which work under simplifying assumptions.