Brain tumor classification from magnetic resonance imaging (MRI) scans performs an essential task in early detection as well as planning of treatment. Convolutional Neural Networks (CNNs) have exhibited outstanding performance in task of image classification, as evidenced by numerous research studies. In this research, we propose a CNN architecture utilizing Efficient Net as the backbone for accurate and efficient brain tumor MRI classification. Efficient Net, a state-of-the-art CNN architecture, has gained significant attention in the field of research for its remarkable combination of superior performance and efficiency. It leverages a compound scaling technique that optimizes model parameters by balancing depth, width, and resolution. This is what makes Efficient Net highly suitable for resource-constrained scenarios, such as medical imaging. This pooling operation summarizes the learned information, enabling more efficient learning. Fully connected layers are then employed to transform the pooled features into a classification decision. Total number of units at final fully connected layer represents number of target classes, in which case would represent different tumor types. The SoftMax activation function applied to final layer, providing possibility scores for each class, indicating the likelihood of the input belonging to a specific tumor category. Experiments show that our proposed method achieves a considerable improvement of 5% in performance metrics, surpassing existing methods in brain tumor classification accuracy.