The breast is a vital organ in the female body, both physically and socially. Breast cancer remains one of the leading causes of cancer-related mortality among women worldwide, making early and accurate diagnosis essential for improving treatment outcomes and survival rates. Deep learning techniques, particularly Convolutional Neural Networks (CNNs), have demonstrated significant potential in breast cancer image classification. However, CNN-based approaches often require a large number of trainable parameters and may not effectively capture both local and global image characteristics. To address these limitations, this study proposes a lightweight hybrid deep learning framework for breast cancer classification using a Kaggle breast cancer dataset. The proposed framework consists of two complementary CNN branches: a local feature extraction network that captures fine-grained cancer characteristics and a global feature extraction network that learns broader contextual information. The extracted features are fused through a concatenation mechanism, followed by feature reduction and classification to generate the final prediction. The whole workflow includes image preprocessing, dual-branch feature extraction, feature fusion, dimensionality reduction, and classification. Experimental results demonstrate the effectiveness of the proposed framework, achieving a classification accuracy of 98.8% on the Kaggle breast cancer dataset. The integration of local and global feature representations enhances the model’s ability to distinguish between benign and malignant breast lesions while maintaining a lightweight architecture with reduced computational complexity. Compared with conventional single-branch CNN models, the proposed approach provides improved feature representation and classification performance. These findings highlight the potential of the proposed framework as an efficient and reliable computer-aided diagnostic tool for early breast cancer detection and clinical decision support.