Ultrasound imaging is a widely used modality in medical diagnostics due to its non-invasive, real-time, and cost-effective nature. However, its clinical utility is often hindered by speckle noise, a multiplicative noise that degrades image quality and complicates subsequent analysis tasks. Traditional denoising methods, such as local adaptive filters and diffusion-based techniques, have demonstrated limited success due to their dependence on predefined parameters and their tendency to introduce artifacts. Recent advancements in deep learning have led to powerful speckle noise reduction techniques, yet their black-box nature limits interpretability. In this study, we propose a novel hybrid approach that integrates deep learning with the Speckle Reducing Anisotropic Diffusion (SRAD) method. Specifically, we employ a neural network to approximate key parameters of the SRAD partial differential equation (PDE), allowing for an adaptive and interpretable edge-preservation speckle reduction. Experimental results demonstrate that our method achieves superior performance compared to conventional filtering techniques and deep-learning-based approaches, offering a robust and explainable solution for ultrasound speckle denoising.