To address challenges such as fast frequency changes and highly dynamic signal environments in frequency-hopping systems, this paper proposes an improved diagonally loaded SMI (Sample Matrix Inversion) algorithm suitable for FPGA (Field-Programmable Gate Array) implementation to enhance the system's anti-interference capability in complex environments. Compared with the traditional SMI algorithm, the improved diagonally loaded SMI algorithm is more effective in handling signal processing demands under low snapshot numbers and complex interference environments, with a lower computational complexity for the diagonal loading factor. This paper briefly introduces the basic principles of the improved diagonally loaded SMI algorithm and the calculation method of the diagonal loading factor, while providing a detailed explanation of its FPGA implementation and performance analysis. Firstly, low-complexity estimation of the diagonal loading factor is achieved using high-level synthesis (Vivado HLS）technology, enabling the optimal weight vector calculation and simplifying the design process. Subsequently, the IP core generated by HLS is packaged and integrated into the project to implement beamforming. Simulation results show that the diagonally loaded SMI algorithm can achieve an interference-to-signal ratio (SIR) improvement of over 70 dB on the FPGA platform, demonstrating significant interference suppression effects. Additionally, the algorithm completes anti-interference weight calculation, and beamforming update within 4734 clock cycles, ensuring fast performance and meeting the real-time processing requirements for frequency-hopping signals with over 10 000 hops per second.