The deep learning algorithms and robotics have emerged in each and every field of human life. Most of the recent innovations in image recognition are dependent on deep learning technology. In this manuscript, we present an intelligent robot that uses this deep learning to detect and classify fire. We have used a combination of Alexnet to detect fire and Imagenet for detecting the type of fire. We found out that the classification accuracy of fire detection goes up to 98.25 %. and classification accuracy of firetype classification was around 92 %. We believe that there is a lot of scope to improve the accuracy in firetype classification but could be a good starting point to design robots as per the class of fire with machine vision.