In order to explore the mechanism of measurement error reduction for a double tipping bucket rain gauge(DTBR),and to further improve its measurement accuracy, a method of combining experiments and numerical simulation was proposed. A high-speed camera was used to get the tipping characteristics during the movement and a dynamic grid and the volume of fluid (VOF) model were used to simulate the overturning process for the upper tipping bucket of the DTBR. The structure design of DTBR was further optimized.The results show that the DTBR relies on the upper bucket to smooth the discharge under natural precipitation intensity at a steady rate of 1.01 g/s injecting into the lower bucket through the middle funnel, which can significantly reduce the influence of precipitation intensity on the measurement error of the rain gauge, and the effect is more obvious under high precipitation intensity. The discharge of the middle funnel under different precipitation intensity remained at a steady rate of 1 g/s (corresponding to precipitation intensity of 1.91mm/min), and as a result, for the future DTBR calibration, the 1.91mm/min precipitation intensity calibration can be considered for the lower tipping bucket, which can meet the error requirements of DTBR. There is an oscillating outflow process of the middle funnel, which can reach a maximum of about 2.g/s. Reducing the central shaft height of the upper bucket can stabilize the outflow of the middle funnel to a certain extent.