In order to accurately evaluate the fracturing stimulation effect of horizontal wells in shale gas reservoirs, it is suggested to establish a model to predict the temperature distribution in the wellbore of horizontal well and an inversion model of DTS data by using MCMC algorithm, and optimize the production profile interpretation process. In this paper, the characteristics of temperature profile of fractured horizontal wells in shale gas reservoirs were analyzed and the main factors affecting the temperature profile were figured out. Finally, the newly established inversion model was applied to the production profile interpretation of one case well in a certain shale gas reservoir. And the following research results were obtained. First, the temperature profile of fractured horizontal wells is in the shape of irregular “saw tooth”, and each “saw tooth” corresponds to an effective hydraulic fracture with fluid inflow. Second, the longer the fracture is, the greater the wellbore temperature drop at the corresponding fracture location is, and the gas flow rate in the fracture is positively correlated with the temperature drop. Third, from the perspective of influence degree, the factors influencing the temperature profile of fractured horizontal wells in shale gas reservoirs are ranked from the strong to the weak as follows: fracture half-length, gas flow rate, permeability of stimulated area, wellbore diameter, fracture conductivity, horizontal inclination angle, and comprehensive thermal conductivity, among which, the first three are main controlling factors. Fourth, the MCMC inversion method is applied to invert the DST temperature data of the case well. The temperature profile predicted in the model is better accordant with the measured DTS profile, and the absolute error of the predicted temperature at different levels of effective hydraulic fractures is less than 0.02 °C. The interpreted gas flow rate of each fracturing stage is closer to the field measurement, and the deviation of the maximum gas flow rate of a single fracturing stage is only 180.35 m3/d. The absolute error between single-well gas production rate and gas production rate measured at the wellhead is less than 3 m3/d, which proves the reliability of this newly developed inversion model.