In order to predict the oil production rate and recoverable reserves in reservoirs developed by SAGD process quickly and accurately, and complete the task of scheme optimization and production control on time, the full life cycle models, like generalized Weng’s model, Rayleigh model, Weibull model, Hubbert model, HCZ model and Maxwell model, were used to carry out corresponding predicting research, then their applicability were assessed. Considering the production capacity calculation formulas of SAGD at different stage of development process, a piecewise regression prediction model on the basis of production capacity formula was proposed to predict oil production rate and recoverable reserves. In order to improve the precision of prediction, the combinatorial prediction model was proposed based on above methods. The results show that the recoverable reserves calculated by combinatorial prediction model are 30.67×10^(4)m^(3), and the recovery factor of related SAGD pad could reach 64.4%, based on the example in this paper. It is concluded that the full-life cycle models and the piecewise regression prediction model on the basis of production capacity formula can be well applied to predict oil production rate and recoverable reserves in reservoirs developed by SAGD technique. These methods are quite simple, practical and feasible. The combinatorial prediction model can further improve the corresponding precision of prediction.