The mind tumors, are the most widely recognized and forceful illness, prompting an exceptionally short future in their most elevated evaluation. In this manner, treatment arranging is a key stage to improve the personal satisfaction of patients. For the most part, different picture methods, for example, Computed Tomography (CT)and ultrasound picture are utilized to assess the tumor in a mind, lung, liver, bosom, prostate… and so forth. Particularly, in this work, CT pictures are utilized to analyze tumor in the cerebrum. Anyway, the colossal measure of information produced by CT examines thwarts manual grouping of tumor versus improvement tumor in a specific time. In any case, it has some impediment (i.e) exact quantitative estimations is accommodated a predetermined number of pictures. Subsequently trusted and programmed order plot is basic to keep the passing rate of a human. The programmed cerebrum tumor characterization is extremely testing errand in extensive spatial and basic changeability of encompassing locale of mind tumor. In this work, programmed mind tumor identification is proposed by utilizing Convolutional Neural Networks (CNN) grouping. The more profound engineering configuration is performed by utilizing little parts. The heaviness of the neuron is given as little. Test results demonstrate that the CNN chronicles rate of 97% precision with low multifaceted nature and contrasted and the all other condition of expressions techniques.