Abstract: The point cloud compression standard “Information Technology Space-Time Graphical Data Coding Part 2: Point Cloud” (i.e. AVS-PCC), developed by the Audio Video Coding Standard Workgroup of China, was promulgated in October 2024. In the octree coding of the geometry compression part in AVS-PCC, the existing method empirically constructs two non-uniform context models based on point cloud density, which is an ineffective approach for the removal of spatial redundancy. To address this issue, a uniform context model based octree coding method is proposed. Firstly, for various types of point clouds with different densities, a uniform strategy is designed to construct state models based on the neighboring geometry information for the current sub-node. Secondly, a finite counter based probability estimation method is proposed to estimate the probabilities of different states in each state model. Finally, a uniform context model is generated guided by the results of the probability estimation. The current sub-node is then compressed using the context-based adaptive binary arithmetic codec. In comparison to the existing method under Common Test Conditions, the proposed method has enhanced the lossy and lossless geometry compression performance of AVS-PCC, with overall average performance gains of 2.5% and 1.2%, respectively.