Owing to the fact that the increasing amount of attention has been focused on numerical weather forecast and climate change research, it is desired that the observation error of upper air temperature with using sounding temperature sensors can be reduced down to 0.1 K. However, the temperature measurement errors of bead thermistor sounding temperature sensors, induced by solar radiation, are on the order of 1 K or more, which is a few orders of magnitude larger than the errors produced by the measurement circuits and digital signal processing systems in radiosondes. Hence, the solar radiation error poses an important bottleneck for improving the measurement accuracy. To tackle this problem, a numerical analysis method is proposed in this research. By employing a computational fluid dynamics (CFD) method, the influences of various solar radiation intensity, sensor angles, and air pressures from sea level to 20 km altitude on temperature measurement accuracy are studied. In this CFD model, the boundary conditions of external convection and solar radiation of the bead thermistor are taken into consideration. The modeling results indicate that solar radiation intensity and altitude are important factors that affect the amplitude of the radiation error. With the elevation increasing from sea level, the solar heating error appears to have an exponential correlation with the altitude, which exhibits a growing slop rate. When the sensor angle is 90o, the radiation error of a bead thermistor sensor probe is minimal. The simulation results are fitted by a Levenberg-Marquardt method and a global optimization method. A correction equation of the radiation error is obtained, where the altitude of the sensor and solar radiation intensity act as two major variables in the equation. In order to verify the equation obtained in this study, an experimental platform for solar radiation error, which includes a low-pressure temperature chamber, a rotation apparatus, an LED-based radiation source, and a wireless communication system, is designed and constructed. It can be found that the solar radiation errors of the bead thermistor based on fluid dynamics numerical calculation are generally consistent with experimental data. The average offset and root mean square error between the correction equation and experimental results are 0.017 K and 0.023 K, respectively, which can demonstrate the accuracies of the computational fluid dynamics method, the Levenberg-Marquardt method and the global optimization method proposed in this research. The methods and techniques introduced in this paper may open the way for correcting the solar radiation errors of the bead thermistor sounding temperature sensors.