Today it is hardly possible to imagine our planet without forests as these zones are associated with wealth in the modern world. They sustain the life cycle of various species and are important for the existence of human beings as well. These forests are often destroyed by such environmental hazards as fires and this results in economical, wildlife, material and human losses. It poses a threat to every object that is close to it. However, the presence of plants and animals only enhances the fire and the rate through which it is transferred. In the case of these forest fires, improvement of measures that enhance early identification of these fires will go a long way in preventing them from spreading to other regions. In this article, we assess existing approaches for the forecast and prevention of forest fires. Last of all, we make a comparison of various machine learning and deep learning models which have been applied for forest fire detection. Furthermore, we discuss advantages and disadvantages that might occur while applying forest fire prediction and prevention systems with the help of a machine and deep learning techniques.