This study investigates data-processing methods and examines the precipitation effect on gravity measurements at the Dali gravity network, established in 2005. High-quality gravity data were collected during four measurement campaigns. To use the gravity data validly, some geophysical corrections must be considered carefully. We first discuss data-processing methods using weighted leastsquares adjustment with the constraint of the absolute gravity datum. Results indicate that the gravity precision can be improved if all absolute gravity data are used as constraints and if calibration functions of relative gravimeters are modeled within the observation function. Using this data-processing scheme, the mean point gravity precision is better than 12 lgal. After determining the best data-processing scheme, we then process the gravity data obtained in the four measurement campaigns, and obtain gravity changes in three time periods. Results show that the gravity has a remarkable change of more than 50 lgal in the first time period from Apr–May of 2005 to Aug–Sept of 2007. To interpret the large gravity change, a mean water mass change(0.6 m in height) is assumed in the ETOPO1 topographic model. Calculations of the precipitation effect on gravity show that it can reach the same order of the observed gravity change. It is regarded as a main source of the remarkable gravity change in the Dali gravity network, suggesting that the precipitation effect on gravity measurements must be considered carefully. This study investigates data-processing methods and examines the precipitation effect on gravity measurements at the Dali gravity network, established in 2005. High-quality gravity data were collected during four measurement campaigns. To use the gravity data validly, some geophysical corrections must be considered carefully. We first discuss data-processing methods using weighted least squares adjustments with the constraint of the absolute gravity datum. Results that the gravity precision can be improved if all absolute gravity data are used as constraints and if calibration functions of relative gravimeters are modeled within using this data-processing scheme, the mean point gravity precision is better than 12 lgal. After determining the best data-processing scheme, we then process the gravity data is in the four measurement campaigns, and obtaining gravity changes in three time periods. Results show that gravity has a remarkable change of mor e than 50 lgal in the first time period from Apr-May of 2005 to Aug-Sept of 2007. To interpret the large gravity change, a mean water mass change (0.6 m in height) is assumed in the ETOPO1 topographic model. Calculations of the precipitation effect on gravity show that it can reach the same order of the observed gravity change. It is as a main source of the remarkable gravity change in the Dali gravity network, suggesting that the precipitation effect on gravity measurements must be considered carefully.