Abstract: Solar radiation is the primary energy source of various physical processes in the natural environment, the basic force of driving formation and evolution of weather and climate. It's one of the most important natural factors in many scientific fields. However, compared with temperature and precipitation, solar radiation data are rather deficient in China, and thus it is necessary to seek a new way to calculate it by using the routine meteorological variables. With the rapid development of the national economy and the continuous expansion of cities, the air quality of some large and medium-sized cities has been gradually worsening. However, these effects are often neglected in radiation model, which may cause great errors in the solar radiation calculation. Therefore, a more precise solar radiation model which regards air pollution index as a dominant factor is built to improve model simulation accuracy. By analyzing the statistical analyses about observations of 23 nationwide sites from 2001 to 2012, the model is established for assessment of daily radiation with nonlinear regression method, which is called DSRM-Y. The model includes three independent variables: Daily temperature range, astronomical percentage of sunshine and air pollution index. Meanwhile, to check the simulation effect, results of DSRM-Y are compared with those of a previous model called DSRM-C. Results show a significant negative correlation between two essential factors of air pollution index and daily solar radiation. In addition, a good fitting effect is expressed in scattering diagram, mean bias error, root mean square error and error analysis of daily total radiation between calculated and measured values. The application result of DSRM-Y at Xining, Shanghai and Kunming indicates, as the air pollution index rises, the daily solar radiation weakens in all three sites. In terms of model comparison, the root mean square error of DSRM-Y is less than that of DSRM-C for all sites, indicating DSRM-Y performs better.