An investigation on the daily photosynthetically active radiation (PAR) with the global solar radiation (R-s) is conducted at 7 surface radiation budget monitoring stations across the Mainland United States by exploiting a 3 years (2009-2011) data achieve. The clearness index, the diffuse fraction and the skylight brightness along with the dew point temperature and the cosine of solar zenith angle are used to generate empirical relationships for predicting PAR from R-s. Records of 2009 and 2010 are employed for model establishment, while records of 2011 are used for validation. The accuracy of the models' predictions is evaluated by four statistics parameters, including the coefficient of determination, the root mean square error, the mean percentage error and the relative standard deviation. Results show that the polynomial model taking the clearness index as main parameter plus the cosine of solar zenith angle has the best performance out of ten proposed models. And the clearness index is capable to be the indicator for PAR prediction, as a substitute of the combination of the diffuse fraction and the skylight brightness. (C) 2014 Elsevier Ltd. All rights reserved.