Electrically heated blankets have been around the market for a long time. They are extremely handy during winters, as they can be used to warm beds. There have been many advances in the control systems to ensure that the desired temperature of the blanket is maintained. The drawback of such systems is its power consumption. Commercially available heating blankets use about 200 watts of power. On average, an adult’s body occupies only about 60 to 70 percent of the total sleeping area of a bed. Existing technologies heat the full blanket irrespective of the users sleeping position and orientation. This results in wastage of power due to inefficient heating of the blanket. In this paper, we present an intelligent heating system that uses Long Short Term Memory (LSTM) to learn the sleeping patterns of the user to predict the future position and orientation of the user to maximize its overall efficiency.