The reasonable and effective exploitation and optimal utilisation of coal resources as well as ensuring the safe operation of main buildings, railways, high-voltage power lines, and other facilities of mining-influenced areas are the main technical requirements of the coal industry. The key problem is to implement certain methods and techniques to acquire deformation information, summarise the deformation laws, and ultimately serve mine production and construction. Considering the surface deformation of mining areas, traditional measurement techniques (automatic/digital levels, theodolites and total stations, etc.) are generally used to measure the well-arranged (several hundreds of) monitoring points repetitively, obtain deformation information, and calculate the deformation parameters. These parameters are used to design protective coal pillars. However, surface deformation is expensive to monitor or measure; it cannot be measured by traditional techniques in several coal mine areas, particularly in mountainous regions. Hence, effective techniques must be employed to obtain deformation data. Although global positioning system real-time kinematic (GPS RTK) technology provides an effective solution to this problem, the traditional GPS RTK with surveying rod can only determine deformation at centimetre level or low positioning precision because the rod is subject to vertical deviation and shaking error; thus, it cannot satisfy the precision requirement of surface deformation monitoring. A new GPS RTK surveying method that involves rod measuring is proposed in this study to address these issues. This method can effectively avoid the effects of the vertical deviation and shaking error of the surveying rod, weaken the impact of multipath error in U direction, and further improve positioning precision. Experiment results show that at 1 s sampling interval, the standard deviations estimated by the proposed method toward N, E, and U directions are 11·4, 8·9, and 4·9 mm for 10 s observation, respectively, and 8·9, 5·1, and 4·0 mm for 20 s observation, respectively. The impact of the major error source was reduced, and positioning precision improved significantly. This particular result provides strong technical support for fast and high-precision mine surface deformation monitoring.