With the rapid development of aerospace technology, the detection of various targets in space has become a crucial research direction for human exploration of the unknown, national security assurance, and the promotion of technological advancements. This paper proposes a ship target detection network based on big data-driven and deep learning techniques. The network integrates feature fusion, regularization enhancement, and performance optimization modules, effectively addressing the challenges of multi-scale, multi-directional target detection and false alarm source interference in complex environments. Experimental results show that the algorithm achieves an accuracy rate exceeding 80% for the recognition of various target types, significantly enhancing detection precision and efficiency. This provides strong technical support for marine resource management and maritime security monitoring.