In recent years, the integration of multimodal data and the application of advanced machine learning techniques have revolutionized the domain of predictive modeling. This research investigates the potential of multimodal data fusion, specifically combining textual and image-based information, using convolutional neural networks (CNNs) to predict house prices accurately. This research paper presents an advanced hybrid convolutional neural network (CNN) architecture designed to predict house prices based on house attributes, including numerical attributes, textual attributes, spatial information, and images. This defines the multimodal attribute of the data used in the study. The proposed model incorporates residual and skip connections to capture complicated and detailed spatial patterns in the input images and improve overall prediction performance. This helps to resolve the vanishing gradient problem that occurs with the deep neural networks. The study demonstrates the effectiveness of the proposed model on a real-world housing dataset, achieving accurate price predictions. The model is evaluated using mean squared error (MSE) as the loss function, and the Adam optimizer is utilized for efficient convergence during training. The research emphasizes the importance of integrating both numerical and visual information for better housing price estimation. Experimental results highlight the capability of advanced CNN to effectively learn features from images, leading to enhanced prediction accuracy in real estate valuation.