Production prediction of shale gas reservoirs during CO2 huff-and-puff based on inflow performance relationship curves

China is endowed with abundant shale gas resources and considerable development potential; however, shale gas production frequently declines rapidly after the initial development stage. CO 2 huff-and-puff has emerged as a promising technique for enhancing shale gas recovery. To address this decline and optimize the huff-and-puff process, reliable production prediction is essential. Although various prediction methods have been proposed, they are unsuitable for direct application to CO 2 huff-and-puff process in shale gas reservoirs. In this study, we developed a productivity prediction model for CO 2 huff-and-puff based on inflow performance relationships (IPR). Using geological data from a representative shale gas reservoir in China, we constructed a multi-stage fractured horizontal well model via numerical simulation, explicitly considering both CO 2 injection and CO 2 -CH 4 competitive adsorption effects. The results indicate that production during the early transient-flow regime is mainly influenced by porosity, huff-and-puff cycle number, Langmuir parameters, and fracture conductivity, whereas during the late-flow regime, porosity and huff-and-puff cycle number dominate. Based on the simulation outcomes, we regressed the key parameters using the Levenberg-Marquardt algorithm to derive predictive expressions for both stages. Validation with independent simulation cases demonstrated that the proposed model achieved mean absolute percentage error ( MAPE ) values of 12.24%, 9.80%, 8.03% for the predicted coefficients a 1 , c 1 and c 2 , respectively, with corresponding Willmott index ( d ) values of 0.74, 0.98, and 0.98. These results indicate satisfactory prediction accuracy and good consistency between the predicted and simulated results. The model thus providing an efficient approach for rapid production forecasting and optimizing CO 2 huff-and-puff strategies.