This research introduces novel approach for image fusion leveraging Deep Convolutional Generative Adversarial Networks (DCGANs) to integrate visual and infrared images into a single, coherent output. Conventional image fusion methods frequently struggle to precisely acquire and merge the complementary information found in multi-modal images, resulting in inferior fused outcomes. On the other hand, our approach utilizes the strong generating ability of DCGANs to acquire a resilient fusion model that maintains essential characteristics from both visual and infrared data. Our approach centres around a generator-discriminator framework, where the generator creates a fused image from a pair of visual and infrared images, and the discriminator assesses the authenticity of the generated image compared to real fused images. The training process is driven by a unique loss function that merges the adversarial loss with an L1 loss, guaranteeing that the generated image looks authentic and preserves its structural resemblance to the ground truth. To evaluate the efficacy of our suggested fusion strategy, we conduct comprehensive tests using a standardized dataset that has been preprocessed to guarantee consistent alignment and normalization. The outcomes of our study exhibit substantial enhancements compared to conventional fusion methods, both in terms of quality and quantity, as indicated by metrics such as Structural-Similarity-Index (SSIM) and Peak Signal-to-Noise Ratio (PSNR). The fused images generated by our model exhibit enhanced detail retention and clarity, effectively merging the salient features of the source images.