Abstract The objective of this work was to estimate the global radiation by simplified models for the state of Mato Grosso, Brazil. The parameterized coefficients of 15 simplified models were regionally calibrated to estimate the daily global radiation, based only on air temperature, using data from 28 automatic weather stations (AWS) of the network of the Brazilian Meteorology Institute, distributed throughout the different biomes of the state of Mato Grosso. The simplified models are mostly derived from the Hargreaves and Bristow & Campbell methods, with different parameterized coefficients to be calibrated. The coefficient of determination (R2), the mean bias error (MBE), the root mean square error (RMSE), and Willmott’s d index were used to evaluate statistical performance. For the recommendation of models per station and/or biome, the models were rated numerically (position values) according to their specific performance in each statistical indicator. The simplified models derived from Bristow & Campbell showed better statistical performances for estimating daily global radiation. The values of the calibrated coefficients of the same model varied greatly among the AWS and biomes. The R2 values ranged from 0.60 to 0.75, indicating a satisfactory result for the obtained calibrations. The Bristow & Campbell model for the Amazon and the Cerrado and the Goodin et al. (1999) model for the Cerrado are recommended, with scattering varying between 1.52 and 4.33 MJ m-2 per day and adjustments greater than 65%.