This paper addresses the challenge of robust face recognition under noisy conditions by evaluating advanced denoising techniques. Specifically, we investigate the effectiveness of Wavelet Transform, Principal Component Analysis (PCA), and Non-Local Means (NLM) denoising methods. The experimental studies were conducted on the Yale face dataset to assess the performance of these techniques. Among the methods tested, NLM denoising exhibited superior performance, effectively preserving edges and fine details while reducing noise. This resulted in enhanced accuracy of face recognition compared to other denoising techniques. Our findings highlight the critical role of advanced denoising in improving the robustness and reliability of face recognition systems in noisy environments.