With the advancement in China’s shale gas industry and the large-scale production of shale gas wells，how to effectively forecast the production rate and recoverable reserves of shale gas wells and has become a crucial subject of oil and gas reservoir engineering. It is also a matter of great concern to shale gas production and management departments. As shale gas is stored in the ultra-tight shale matrix and the secondary fracture system in the adsorption state and the free gas state，respectively，it can be polluted by mud during the drilling，completion，logging，and fracturing. Therefore，the shale gas wells that have been put into production show obvious independence，diversity，and complexity. More importantly，shale gas wells are brought into production by desorption after pressure drops and energy consumption featured by the decreasing production at the moment it starts. Consequently，the production decline method has become the first choice for evaluation.In this paper，the generalized single peak cycle model proposed by Chen Yuanqian et al.is simplified into a pan exponential decline model（PEDM）that can forecast the production rate and recoverable reserves of shale gas wells. The decline exponent m ranges from 0 to 1. At m=1，it is the famous exponential decline model. At m=0.5，it is the 0.5-PEDM of practical value. For specific unconventional shale gas wells，it is necessary to use production data to determine the values of a，c，and m by linear iterative trial and error method in actual forecasting. The application of production data from two wells in the Marcellus shale gas reservoir，Pennsylvania，the US，elicits the effectiveness of the PEDM in forecasting the production rate and recoverable reserves of shale gas wells.