Identification of inner heat source based on surface temperature measurement can find many applications such as in thermographic nondestructive defection, in microbial fermentation and in ammunition and food storage fields. Genetic algorithm was used to study the identification problem of the heat source term. The factors affecting the identification results were analyzed. Numerical examples proved that the method can provide a good solution to the parameter identification problem of inner point heat source. The identification results of position and strength are accurate when multiple point heat sources are identified simultaneously. Only the identification result of the position has a deviation when the point heat source is far from the unique temperature measurement boundary. The identification results of positions and strengths of point heat sources are accurate when the standard deviation of the temperature measurement error does not exceed 1.0 ℃.