To address the challenge of threat assessment for multi-space non-cooperative targets，a multi-space non-cooperative target threat assessment method based on the dynamic Bayesian network （DBN） and the technique for order preference by similarity to ideal solution （TOPSIS） is proposed.First，the key threat factors of spatial non-cooperative targets are analyzed and extracted from a mission-oriented perspective，and a DBN model is constructed to represent their dynamic interrelationships.Meanwhile，an intent recognition module is incorporated into the threat assessment framework to enhance the accuracy and rationality of threat level estimation for individual targets.Second，the TOPSIS method is adopted to quantify and rank the threat levels of multi-space non-cooperative targets.Finally，simulations and analyses are conducted for the target threat assessment under representative scenarios to validate the effectiveness and applicability of the proposed threat assessment approach.