Brain tumor detection is an important application in recent days. The medical problems are severe if tumor is identified at the later stage. Hence diagnosis is necessary at the earliest. MRI is the current technology which enables the detection, diagnosis and evaluation. In this work, the images obtained through MRI are segmented and then fed to a model known as Pulse coupled neural network for detecting the presence of tumor in the brain image. The physician could seek the help of this model if the input MRI brain images are more in number and the network would help the physician to save time for further analysis. The work also utilizes back propagation network for classification. Both the networks are less complex in nature and hence the processing of MRI brain images is very simple. The network classifies the input images as normal and tumor containing. The tumor may be benign and malignant and it needs medical support for further classification.