Order promising is a critical capability helping companies win competitive advantage while considering customer orders. Available-to-Promise (ATP) systems as tools for order promising have become popular in practice as well as research. Motivated by its wide applications and the limited research in related areas, this research studies ATP in make-to-stock (MTS) and make-to-order (MTO) production systems under uncertain circumstances by adopting revenue management principles. The first part of the thesis studies the order acceptance and order promising problems that frequently arise in make-to-stock manufacturing systems with multiple demand classes in a finite time horizon. Demands in different periods are assumed to be random variables and independent of one another. Replenishments of inventory usually deviate from the scheduled quantities. The order acceptance model is designed to maximize the expected net profit over the planning horizon by determining the fraction of the demand that is likely to be fulfilled from the inventory. Next, the model is extended to an order promising system where expected future replenishments are also included to deal with incoming orders and order acceptance and delivery dates are determined as appropriate. The characteristics of the optimal order acceptance policy structure are studied to help develop the appropriate heuristic algorithms based on stochastic approximation. Finally, computational experiments are used to illustrate the effectiveness and efficiency of the proposed approaches, and evaluate the applicability of the proposed models. The second part of the thesis considers order promising problems in a make-to-order operation environment. A stochastic ATP system is proposed to allocate critical raw materials and production capacity to uncertain demands in different customer classes. An ATP allocation model and two delivery date assignment (DDA) models are developed to tackle order acceptance and scheduling decisions respectively. The model involves tradeoffs between committing resources to order with lower profit margins, and reserving resources for more profitable. Tradeoffs among the revenue, the holding cost and the tardiness cost are also considered while modeling delivery date assignment. Numerical tests are carried out to investigate the impact of different factors on the profits generated from two DDA models and assess the performance of the stochastic ATP approach. Results from experiments indicate that the models proposed in this thesis do provide profitable opportunities for improving the current ATP performance in both of MTS and MTO manufacturing scenarios.