Abstract: Focusing on the energy-saving solar greenhouse without a rear wall on the northern slope of the Tian Shan Mountains in Xinjiang Uygur Autonomous Region, this study aims to address the issue of excessive and difficult-to-determine parameters in the establishment of greenhouse temperature prediction models. Based on the single crop coefficient method for estimating crop evapotranspiration recommended by FAO-56, a temperature prediction mechanistic model of the solar greenhouse environment in winter was constructed. A comparison with the evapotranspiration estimating model based on the Penman-Monteith method revealed that the predicted indoor greenhouse temperatures were consistent with the measured values. The temperature prediction model based the FAO-56 single crop coefficient method achieved a correlation coefficient R2 of 0.85, while the model based on the Penman-Monteith method achieved an R2 of 0.82. Therefore, with the higher accuracy, a simpler model structure and fewer parameters, the former temperature prediction model provides theoretical support and model validation for the goal of reducing energy consumption of the greenhouse environment control in winter.