Dual-well steam assisted gravity drainage(SAGD) has significant potential for extra-heavy oil recovery.China is conducting two dual-well SAGD pilot projects in the Fengcheng extra-heavy oil reservoir.Quick,direct predictions of the oil production rate by algebraic models rather than complex numerical models are of great importance for designing and adjusting the SAGD operations.A low-pressure scaled physical simulation was previously used to develop two separate theoretical models corresponding to the two different growth stages observed in the SAGD steam chambers,which are the steam chamber rising stage and the steam chamber spreading stage.A high-pressure scaled model experiment is presented here for one dual-well SAGD pattern to further improve the prediction models to reasonably predict oil production rates for full production.Parameters that significantly affect the oil recovery during SAGD were scaled for the model size based on the reservoir characteristics of the Fengcheng reservoir in China.Experimental results show the relationship between the evolution of the steam chamber and the oil production rate during the entire production stage.High-pressure scaled model test was used to improve the gravity drainage models by modifying empirical factors for the rising model and the depletion model.A new division of the SAGD production regime was developed based on the relationship between the oil production rate and the evolution of steam chamber.A method was developed to couple the rising and depletion models to predict oil production rates during the SAGD production,especially during the transition period.The method was validated with experiment data and field data from the literature.The model was then used to predict the oil production rate in the Fengcheng reservoir in China and the Athabasca reservoir in Canada. Dual-well steam assisted gravity drainage (SAGD) has significant potential for extra-heavy oil recovery. China is conducting two dual-well SAGD pilot projects in the Fengcheng extra-heavy oil reservoir. Quick, direct predictions of the oil production rate by algebraic models rather than complex numerical models are of great importance for designing and adjusting the SAGD operations. A low-pressure scaled physical simulation was previously used to develop two separate theoretical models corresponding to the two different growth stages observed in the SAGD steam chambers, which are the steam chamber rising stage and the steam chamber spreading stage. A high-pressure scaled model experiment is presented for one dual-well SAGD pattern to further improve the prediction models to reasonably predict oil production rates for full production. Parameters that significantly affect the oil recovery during SAGD were scaled for the model size based on the reservoir characteristics of the Fengcheng rese rvoir in China. Experimental results show the relationship between the evolution of the steam chamber and the oil production rate during the entire production stage. High-pressure scaled model test was used to improve the gravity drainage models by modifying empirical factors for the rising model and the depletion model. A new division of the SAGD production regime was developed based on the relationship between the oil production rate and the evolution of steam chamber. A method was developed to couple the rising and depletion models to predict oil production rates during the SAGD production , especially during the transition period. The method was validated with experiment data and field data from the literature. The model was then used to predict the oil production rate in the Fengcheng reservoir in China and the Athabasca reservoir in Canada.