Brain tumour detection is an accepted testing task in the beginning phases of life. Yet, presently it has progressed with different AI calculations. Along these lines, to identify the brain tumour of a patient we consider the data of patients like MRI pictures of a patient's brain. Here our concern is to spot if the tumour is available in persisting’s brain. It's crucial to identify the tumour at the beginning level for a sound lifetime of a patient. During this task, we gauge the seriousness utilising Convolutional Neural Network calculation which gives us precise outcomes. Tumour division is a pivotal and burdensome assignment inside the territory of clinical picture preparation as a human-helped manual characterization may bring about erroneous expectations and findings. We suggested using convolutional neural organisation to distinguish tumours from 2D appealing reverberation brain images (MRI). The exploratory analysis was performed on a dataset that included tumours of various sizes, regions, forms, and picture forces. We used convolutional neural networks, which are implemented using Keras and Tensorflow because they respect a much better display than standard ones. CNN achieved an accuracy of 83.54% in our research. The recently developed CNN design may be used as a productive decision-aid device for radiologists in clinical diagnostics because of its high speculation capacity and quickness.