At this point, the conventional smoke detection often requires many smoke sensors but they have weakness in wide coverage area and low response time. In order to overcome these shortcomings, this paper presents a method based on image processing techniques, capable to detect the smoke from video taken from camera. The proposed detection method consists of the following steps: slow motion detection, smoke color detection, inverse edge detection, AND operator and then classification phase. This will provide early warnings such as fire, thus reducing economic losses and casualties. In addition, it will help to improve the rate of smoke detection, as well as reducing the false detection rate of other suspected object.