Image denoising is a vital process in improving the visual quality by eliminating unnecessary noise from images, to restore the original clarity images. Conventional methods often assume Gaussian-distributed noise, which may not represent accurately for real-world noise complexities. This proposed research focuses on optimizing blind noise, prevalent in real-world images using deep learning techniques. The denoising model learns the features directly from the noisy samples by capturing the noise distribution within the image. This study employs the deep learning fusion-based method by integrating the noise prioritization using the Pyramid Real Image Denoising Network (PRID Net), multi-scale feature extraction, and Residual Encoder-Decoder Network (RED net) to reduce the blind noise. Both PRID network and RED network use the convolution neural network to efficiently lower the noise in order to enhance the noisy image’s quality. The experimental analysis compares the proposed model with the individual PRID and RED network models in terms of objective measurements such as the Structural Similarity Index (SSIM), Mean Squared Error (MSE), and Signal-to-Noise Ratio (SNR) on various datasets such as RENOIR, SIDD, and NIND datasets.