The targets cannot be accurately extracted in the existing infrared image for environmental factors and background interference.To solve this problem,an extraction algorithm with high precision infrared target separation was proposed based on the radiation characteristics of the object.The amount of radiation and the probability distribution were calculated by using radiation distribution characteristics of the object.The amount of radiation was converted to gradation gray level in the last.And combined with the characteristics of gray difference,the high-precision separation of the goal was achieved.Experimental results show that this algorithm can obtain more accurate target image compared with other algorithms.And it has the advantages of strong anti-interference ability and low time-consuming.