In this paper, we present a novel distributed sequential allocation mechanism for optimizing investments in unmanned systems, framed as a multi-agent dynamic planning problem. The core contribution lies in an advanced algorithm that integrates multi-agent parallel computing for global optimization with single-agent sequential allocation for local refinement. This hybrid approach ensures both optimality and polynomial-time complexity, effectively addressing the challenges of multi-field investment with uncertain costs and rewards. By employing sophisticated optimization techniques, our algorithm dynamically adjusts investment strategies based on the real-time data. Simulation results in typical scenarios demonstrate the algorithm’s superiority over benchmark methods, offering significantly enhanced investment solutions tailored to the unique requirements of unmanned systems. Our method not only improves investment efficiency and effectiveness, but also provides a robust and adaptable solution for the dynamic and uncertain nature of unmanned systems investment portfolios, thereby ensuring sustained performance and strategic advantage.