Forest fires are often enlarged over extended periods of time, but if they are detected through image-based real-time monitoring, this would shorten the response time of fire fighters, thereby preventing great damage. When a forest fire occurs, it is often difficult to access the area and to accurately identify the location of its occurrence through monitoring images only. To overcome these problems, this paper proposes a method of detecting forest fires using infrared cameras and estimating the location of the forest fire by applying the space forward intersection method. Additionally, practicality verification was conducted through an outdoor ignition experiment, and the measurement accuracy of distance was compared.