When the radiation source threat assessment is treated as a multi-attribute decision problem, due to the complexity of the operational environment, the scouting party cannot obtain complete information of the enemy radiation source, and the attribute value and attribute weight of each radiation source target are usually not determined values. There is a certain degree of uncertainty, so it is generally expressed by the interval number; in addition, the Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS) method only considers the Euclidean distance between the indicators, and cannot reflect the correlation between the indicators. Aiming at these problems, this paper introduces interval number theory and grey relational analysis(GRA) to extend the traditional TOPSIS method. An interval multi-attribute decision making method based on GRA and TOPSIS is proposed to solve the radiation source threat assessment problem under uncertain conditions. This method constructed Euclidean distance based on interval number and grey relational degree based on interval number, constructed decision information system about battlefield situation by combining them, and realized quantitative evaluation of radiation source threat degree by constructing new comprehensive proximity degree. Finally, the simulation results show that the proposed method can effectively solve the problem of radiation source threat sequencing under uncertain conditions in actual combat, which is helpful for the scouting party to further grasp the battlefield situation.