Wildfires are one of the most destructive disasters that have the ability of causing enormous loss to life and nature. Moreover, with its capability to spread abruptly over huge sectors of land, the loss to mankind is unimaginable. Global warming around the world has led to increase in the wildfires, therefore demands immediate attention of the concerned organizations. Conspicuously, this research aims at predicting the forest fires to minimize the loss and immediate actions in the direction of safety. Specifically, this research proposes an energy efficient IoT framework backed by fog-cloud computing technology for early prediction of wildfires. Initially, Jaccard similarity analysis is used to determine the redundant data acquired from IoT devices in real-time. This data is analyzed at fog computing layer and reduces multi-dimensional data to single value termed as Vulnerability Index. Finally, Artificial Neural Network is used to predict the vulnerability on forest region based on Wildfire Causing Parameters. ANN model is appended with Self-Organized mapping technique for effective visualization of geographical region with respect to wildfire vulnerability. Implementation simulation is performed over different datasets. Results are compared with several state-of-the-art techniques for overall performance estimation.