Abstract: To the problem of low efficiency and high energy consumption for the traditional forest fire prevention methods,this article discussed the significance of forest fire prevention based on Sea computing,and put forward the implementation of forest fire recognition algorithm based on Sea computing.Firstly,used Qt/Embedded to design according to the characteristics of the sensor node.Then,used 24-bits additive color model and Sobel operator to complete the decision of color and region based on the features of forest fire.Finally,the results were fed back to the sink node.This algorithm could automatically analysis image captured by camera and complete the decision of fire in sensor node.The experiments showed that,compared to the traditional methods of forest fire prevention,this algorithm not only completed the instant detection and recognition of forest fire,but also achieved high accuracy rate and low time-consuming at the same time.