The occurrence of fire should be reported by a fire alarm at the earliest. With the advent of convolutional neural networks, this has become more reliable and advanced. The detection of fire conditions as early as possible is a prerequisite of a Fire Detection System, and to give sufficient time for effective counter actions by Automated Fire personnel. Fire detection and forecast by conventional fire detection systems was achieved by using the by-products of fire such as smoke, temperature and flame which take a considerable amount of time to produce a level of fire which will trigger the smoke sensors and heat sensors. In this project, the proposed system is building a deep learning solution using convolutional neural networks. The proposed system is fine-tuned to stabilize the efficiency and accuracy by keeping in mind the nature of the target problem and fire data. The effectiveness of the proposed system is revealed and validated by the experimental results on benchmark fire datasets.