Abstract: In gold exploration relying on unmanned aerial vehicles (UAVs) in mountainous river valley areas, traditional flight route planning algorithms have problems such as a large amount of waypoint data, severely jagged flight trajectories, and difficulty in directly adapting to the flight control system. To address these issues, by taking the Daqiao Gold Deposit as an example, an intelligent path planning algorithm integrating geological fluid potential field and adaptive trajectory sparsification (Geo−Flow−RDP) was proposed. This algorithm ensured the coverage of geological targets within the field of view while achieving deep compression of waypoints, and it underwent multiple simulation tests on the digital elevation model. Geo−Flow−RDP extracted characteristic points from the initial flight route and performed sparse processing. By setting a distance threshold, it minimized redundant waypoints while maintaining the macroscopic geometric features of the route, effectively solving the key technical bottleneck of converting "theoretical algorithm-based flight routes" to "engineering-oriented actual flight routes" in actual geological exploration. After sparse processing by Geo−Flow−RDP, the number of key waypoints for a single flight route was reduced from 2 841 in the traditional Dijkstra algorithm to 23. The data compression rate was as high as 99.2 %, significantly reducing the link load in the weak network environment in the field. Geo−Flow−RDP provided a low-cost and highly reliable gold exploration scheme relying on UAVs for the new round of exploration breakthrough strategies in mountainous river valley areas.