In the future urban environment, unmanned systems will become an important combat force by virtue of their long-range manoeuvrability, high mobility and low-risk attributes. However, the instability of the physical environment in urban battlefields directly affects the autonomous performance of unmanned systems, and the dynamics and uncertainty of urban environments pose additional challenges for unmanned systems to perform their missions, requiring the algorithms in the OODA loop(Observe, Orientate, Decide, Act) to be able to respond quickly to adapt to the rapid changes in the environment . In this paper, through in-depth analysis of the overall impact of urban battlefield physical environment elements on the performance of unmanned systems and their autonomous algorithms, we explore the corresponding technical strategies and algorithmic optimisation paths, and put forward the following strategies to enhance the comprehensive adaptability and intelligence of unmanned systems: firstly, the multi-sensor integration and communication redundancy design at the hardware level, secondly, the algorithmic innovation and the construction of collaborative perception network at the software level, and thirdly, the modularity and scalability of the system design, as well as the construction of the experimental environment.