Unmanned aerial vehicles (UAVs) have become the core infrastructure of smart cities because of their fast, flexible, and strong environmental adaptability. However, signal occlusion caused by the urban canyon effect will seriously affect its global navigation satellite system (GNSS) reliability. The fusion algorithm represented by Kalman Filter cannot meet the real-time and stability requirements of UAV high maneuvering flight positioning due to its high complexity. In this article, integrated GNSS, inertial navigation, and wireless base station navigation, a 3-D UAV positioning method called GIW-UP, based on information geometry, is proposed. It converts the information of various types of navigation sources into probability density functions, and then the fusion is realized from the perspective of information probability. Given the differences in the information output time and navigation parameters of various navigation sources, the proposed GIW-UP method is compared with the least squares (LS) method, the unscented Kalman filter (UKF) method, and the neural network-based multisensor two-stage fusion (MTFA) method in three aspects: stability, convergence speed, and computational complexity. The results show that the GIW-UP can effectively reduce the fusion computational complexity and improve positioning stability.