Forest fire is a major natural disaster in the world. This has provided major challenges for researches to study and predict on ecology imbalances. The high severity forest fire causes most loss to the vegetation and atmosphere by producing chemicals like Carbon to air. So the effects of fire have to be detected within forest region in order to detect the loss. This detection helps to forest department to make a decision for the re-use of the forest zone. Currently satellites are providing continuous monitoring of the fire with videos. But in real time videos cannot provide the data of fire for the loss into the ecology. So proposed work gives a framework where fire occurred place data is collected and stored to do analysis and prediction of fire to ecology. This paper presents comprehensive analysis based on air quality degradation, vegetation loss and high lights need for effective land management strategies. Different supervised learning algorithms are used to predict effect on the atmosphere. The prediction on the effects air and vegetation loss is obtained with the good accuracy from the work