The process of combining images from multiple medical imaging sources without any distortion or loss of information is known as medical image fusion. It improves the usefulness of medical imaging for the diagnosis and treatment of medical issues by maintaining every feature in the fused image. Recent years have seen the introduction of many image fusion techniques that show the significant progress in the field of medical diagnosis. However, the fusion performance of these contemporary methods is prone to distortion, noise, and blurring. To address these drawbacks of the current methods, this work presents a novel fusion methodology for a visually pleasing appearance. Using Gaussian curvature and guided filters, a new decomposition technique with an edge-preserving smoother is firstly described. The primary objective of this approach is to preserve edge haloes while recovering the structural information. The "weighted average" technique is used to merge source images based on the weights that were computed based on significant edge details. This technique can enhance contrast and highlight significant details in the fused image. We evaluate our proposed methodology on multiple pairs of publically available medical imaging datasets. The quantitative evaluation indicates that the suggested fusion strategy for multimodal image fusion improves the average IE by 5.8, MI by 34.8%, MSSIM by 31%, and Q_AB/F by 40% over the current methods, making it appropriate for usage in the medical area for correct diagnosis.