Dynamic pricing is commonly adopted in selling perishable products with fixed stock and selling horizon. Because of the technological change and the volatility of customers’ tastes, the probabilistic demand model commonly used in practice cannot be estimated accurately. This paper considers a retailer’s dynamic pricing problem with little information on demand, and proposes a new approach to deal with this problem. We model the uncertain demand as a hybrid variable which describes the quantities with fuzziness and randomness. The dynamic pricing problem is formulated as three types of hybrid programming models—expected revenue maximization model, -revenue maximization model and chance maximization model—to meet different goals. To solve the proposed models, a hybrid intelligent algorithm is designed by combining hybrid simulations and genetic algorithm. Numerical examples are also presented to illustrate the modeling idea and to show the effectiveness of the proposed algorithm.