In this research, deep learning models such as Convolutional Neural Networks and Recurrent Neural networks are used in diagnosing brain cancer using MRI images. The collection comprised of 2300 MRI images with a variety from different ages and levels for the various stages of brain cancer, each processed using cutting edge preprocessing that improved image quality while making them compatible to ensure simplicity in their application. This study combines CNN architectures like Inception, VGG16, and RestNet with RNNs to include both spatial and temporal information across MRI slices. The Inception with RNN model achieved the best accuracy at 97.55%, followed by VGG16 with RNN (95.60%), and ResNet with RNN (93.40%). Each model’s effectiveness has been evaluated using the performance metrics such as precision, recall, F1 score, and AUC-ROC values. All models were correctly able to identify tumor locations, as shown by confusion matrices and the Inception with RNN model exhibiting minimum misclassifications. These findings also illustrate that the Inception with RNN model has excellent diagnostic ability and suggests it may be useful in optimizing brain cancer diagnosis and prognosis. The proposed research is a powerful framework for applying deep learning to medical imaging, allowing us to make further progress in early detection and treatment of brain cancer.