Abstract: Abstract: Complete and accurate global solar radiation data at a specific region are highly crucial to regional crop modeling, evapotranspiration estimation, irrigation system development and utilization of solar energy resources. In order to improve the utilization efficiency of facility agriculture technology on radiation resources, daily climatic data in 10 meteorological stations in Northwest China from 1993 to 2016 were used to calculate daily global solar radiation. The applicability of 6 solar radiation models based on sunshine (Ångström-Prescott, Ogelman, Bahel, Louche, Almorox-Hontoria and Glower-McCulloch model) and 3 solar radiation models based on temperature (Hargreaves-Samani, Annandale and Bristow-Campbell model) was evaluated using the coefficient of determination (R2), root mean square error (RMSE), normalized root mean square error (nRMSE), Nash-Sutcliffe coefficient (NS), mean absolute error (MAE) and relative error (RE). In addition, the parameters of Bristow-Campbell (B-C) model were calibrated based on least square method using the daily meteorological data from 1993 to 2010, and the daily meteorological data from 2011 to 2016 were adopted to evaluate the applicability of B-C model. The results showed that the estimated and measured daily global solar radiation had statistically significant correlations (P<0.01) for all the 9 models at the 10 stations. Generally, the models based on sunshine (with R2 ranging from 0.875 to 0.954 for the 10 stations) were more accurate to estimate daily global solar radiation than the models based on temperature (with R2 from 0.652 to 0.813). In Northwest China, the Bahel model showed the best performance for daily global solar radiation estimation among the 6 models based on sunshine, followed by Ogelman model and Glower-McCulloch model, with average RMSE of 2.282, 2.309 and 2.313 MJ/(m2•d), average nRMSE of 14.0%, 14.2% and 14.2%, average MAE of 1.666, 1.701 and 1.697 MJ/(m2•d), and average NS of 0.905, 0.903 and 0.902, respectively. The best model based on temperature for estimation of daily global solar radiation was the calibrated B-C model, followed by Hargreaves-Samani model and Annandale model, with average RMSE of 3.819, 4.167 and 4.368 MJ/(m2•d), average nRMSE of 23.3%, 25.5% and 26.6%, average MAE of 2.680, 3.068 and 3.378 MJ/(m2•d), and average NS of 0.741, 0.688 and 0.656 respectively. On the monthly scale, the estimated and measured monthly average daily global solar radiation had good agreement, and the monthly average daily global solar radiation all increased from January to June and decreased from July to December. The Bahel model also showed the best performance for monthly average daily global solar radiation estimation among the 6 models based on sunshine, followed by Ogelman model and Glower-McCulloch model, with average RE of 5.350%, 5.375% and 5.679%, respectively. The calibrated B-C model had the best estimation accuracy among the models based on temperature, with average RE of 4.254%. It meant that the B-C model was more accurate than the Bahel model for global solar radiation estimation in monthly scale. Overall, the Bahel model is recommended to estimate daily global solar radiation when the sunshine hours are available, and the calibrated B-C model is recommended to estimate daily global solar radiation in Northwest China when only temperature data are available. This paper provides valuable information for global solar radiation estimation method in Northwest China, and also provides the theoretical basis and technical support for accurate estimation of daily global solar radiation.