This paper presents a radar target detection method based on Gaussian denoising to enhance the performance of radar systems in noisy environments, particularly in low signal-to-noise ratio (SNR) and long-distance scenarios. Radar data is preprocessed by applying a combination of frequency domain low-pass filtering and Gaussian filtering to suppress noise. This denoising technique is integrated into the RODNet detection model, which is designed to improve target detection accuracy under complex environmental conditions. Experimental evaluations are conducted in both low SNR environments and high SNR conditions with weak targets, comparing the performance before and after denoising. The results show that Gaussian denoising significantly improves target detection accuracy, with the average precision (AP) increasing by approximately 3–5 percentage points. The proposed method enhances radar detection in environments with substantial noise interference, confirming its potential for improving detection robustness. Future research will focus on further optimizing the denoising algorithm’s adaptability and real-time performance to handle more complex scenarios effectively.