Brain tumor classification using MRI scans is a critical task in medical imaging, essential for accurate decision-making and treatment planning. This study leverages a combined dataset of 7023 brain MRI images sourced from Figshare, the SARTAJ dataset, and the Br35H dataset, with images categorized into four classes: Meningioma, Glioma, pituitary tumor, and no tumor. We used the VGG16, VGG19, ResNet-50, and DenseNet169 architectures for classification. Explainable AI (XAI) techniques, such as Grad-CAM, Grad-CAM++, and LIME, were used to generate visual explanations of the model predictions to ensure model interpretability. The results demonstrate that DenseNet169 along with integrating XAI techniques improve both model transparency and diagnostic trust. This work highlights the advantage of integrating classification models and interpretability tools to enable a more inviting AI adoption in healthcare.