Efficient autonomous mission planning for unmanned clusters in digital environments poses ongoing challenges. In response, this paper proposes a task planning model for unmanned cluster systems, based on data fusion and knowledge inference. This paper establishes a semantic framework for task allocation in unmanned drone clusters, utilizing knowledge representation within knowledge graphs and enriching the model through knowledge graph inference techniques. The proposed method meticulously maps the transition of unmanned cluster control and mission planning from physical to virtual spaces. The proposed semantic model of the intelligent architecture based on ontology utilizes knowledge representation and rule inference to enhance cognitive reasoning services for task planning in complex cluster scenarios. This enhancement significantly improves the autonomy of the task execution process. Additionally, the paper includes experimental comparisons between this method and existing advanced algorithms, providing a detailed analysis of its strengths and weaknesses.