As personalized wireless communication grows, cellular-connected UAVs are gradually becoming a crucial means for rapidly constructing wireless communication networks. Due to the impact of complex urban environments on wireless communication, determining the optimal flight path for UAVs poses significant challenges. Existing methods for evaluating the communication performance of UAV paths typically rely on two-dimensional line-of-sight (LOS) probability models with fixed flight heights. This paper considers that UAVs can improve communication quality by adjusting their flight altitude and proposes a three-dimensional LOS probability model to construct a three-dimensional map of the Signal to Interference plus Noise Ratio (SINR), and based on communication quality threshold discrimination, obtains a binary 3D-SINR map. Then, based on this map, using communication interruption rate, duration of interruption, and number of service base station switches as communication quality evaluation indicators, an UAV three-dimensional path planning model is established. Finally, an improved heuristic method is used to solve the UAV flight path. Simulation experiments demonstrate the effectiveness of generating a three-dimensional SINR map in urban communication scenarios. UAV flight paths obtained based on this map have better communication connection quality and shorter path lengths compared to existing methods.