Determining dynamic prices of rooms in hotels is a challenging problem, a subset of Revenue Management (RM). In this paper, the RM task is formulated as a Sequential Decision problem. The utility of the agent’s actions does not depend on single decisions, but on the entire sequence during the booking window, typically months before check-in. Decisions are made with uncertainty and information which is progressively revealed in time. Every price change for future days will actively influence the arrival of future reservations through customer elasticity. Therefore, the initial forecasts and the optimal prices for the future need to be continuously revised. We demonstrate that a price-control system using Sequential Decision control can produce significant and robust results even in the (very frequent) situation of grossly inaccurate demand forecasts and estimates of elasticity. The expected revenue is robust to estimation errors, provided that one starts selling in advance and that prices are revised frequently before the check-in date. In addition, specific constraints can be easily incorporated (like increasing prices) and the tradeoff between expected revenue and risk can be adjusted to the management context.