3D semantic occupancy has rapidly emerged as a research focus in autonomous driving and robotics environment perception, owing to its capability to provide more realistic geometric perception and closer integration with downstream tasks. By performing occupancy prediction of the 3D space in the environment, it effectively enhances the ability and robustness of scene understanding. However, existing occupancy prediction tasks are predominantly modeled using voxel or point cloud-based approaches: voxel-based network structures often suffer from spatial information loss due to the voxelization process, while point cloud-based methods, though better at preserving spatial location information, face limitations in representing volumetric structural details. To address this issue, we propose a dual-modal prediction method based on 3D Gaussian sets and sparse points, which balances spatial location and volumetric structural information to achieve higher accuracy in semantic occupancy prediction. Specifically, our method adopts a Transformer-based architecture, taking 3D Gaussian sets, sparse points, and queries as inputs. Through the multi-layer structure of the Transformer, enhanced queries and 3D Gaussian sets jointly contribute to semantic occupancy prediction, and an adaptive fusion mechanism integrates the semantic outputs of both modalities to generate final prediction results. Additionally, to further improve accuracy, we dynamically refine the point cloud at each layer to enable more precise location information. Experiments conducted on the Occ3D-nuScenes dataset demonstrate superior performance of the proposed method on IoU-based metrics.