Forecasting production in shale gas reservoirs accurately has been of growing interest in the industry. Horizontal wells with multiple fractures are commonly utilized to develop shale reservoirs, which indicates that the dominant flow regime observed will be linear flow for several years. Until now, it has been widely recognized that the rate-transient data analysis is the most efficient approach to estimate rate, where it appears as a straight line on the plot of normalized pressure vs. square root of time in linear flow. However, the traditional square-root-of-time plot may result in overestimation of reservoir properties and will not allow us to forecast production with confidence in shale gas reservoirs. In this paper, a new analytical methodology is put forward to analyze the rate-transient data from fractured wells in shale gas reservoirs producing at a constant flowing-pressure, which incorporates both slip flow/Knudsen diffusion of bulk gas and surface diffusion of adsorbed gas directly into the model. These flow mechanisms cannot be well described by traditional models. Depending on flow discrepancies from conventional reservoirs, the modified pseudo-pressure and pseudo-time equations to account for these critical transport mechanisms are constructed. In addition, a new procedure for rate-transient data analysis applying the proposed model is presented in details, which is reliable and easy to utilize. The novel approach is validated against numerically simulated cases and field observations. Comparisons between the new approach and traditional method are conducted by a number of test cases. The results demonstrate that the newly developed model dramatically eliminates the inaccuracy of production forecast and provides a more reliable estimated ultimate recovery (EUR). This work should provide an efficient guidance to assist analysts in evaluating hydrocarbon production accurately in shale gas reservoirs. (C) 2018 Elsevier Ltd. All rights reserved.