According to the National Cancer Institute (NIH), over 300,000 cases of brain tumors are reported worldwide each year. Brain tumors are characterized by the abnormal growth of cells within various regions of the brain. This organ possesses a complex anatomy, with distinct parts responsible for diverse neurological functions. Tumors can develop in various locations, including the skull, brain stem, sinuses, and nasal cavity. With over 130 known types, brain tumors are classified based on their tissue of origin and location within the brain. Early detection of brain tumors is crucial for effective treatment and improved patient outcomes. Convolutional neural networks (CNNs), especially architectures such as residual networks (ResNet), very deep convolutional networks (VGGNet), and MobileNet, have shown promise in addressing this challenge. This study proposes an Ensemble Model that combines the strengths of these architectures to enhance brain tumor detection accuracy. The proposed model utilizes the combined features of ResNet, VGGNet, and MobileNet to achieve superior performance. An Ensemble Model combining these architectures demonstrates exceptional accuracy, achieving a remarkable 99.7% accuracy in detecting brain tumor cells. Notably, the ROC curve demonstrates an AUC of 1.0 for all four classes, indicating perfect accuracy. The exceptional performance of the Ensemble Model reinforces its reliability in classifying brain tumors and provides a comprehensive and precise solution. This systematic approach demonstrates how deep learning models can significantly enhance diagnostic capabilities for brain tumors and improve clinical decision-making.