Abstract: The assessment model based on cloud theory and composite weighting is introduced to address the uncertainty and subjectiveness in evaluation index estimates. The subjective weights are decided with order relation method and objective weights are defined using the method of Criteria importance through intercriteria correlation. The optimization strategy combines the expert experience with data-based information according to the game theory and calculate the composite weights. The weighted deviation of the trainee’s action from the expected operation action is estimated by means of the cloud theory. This design allows transformation between quantitative scores and qualitative remarks and the ambiguity of remarks is avoided. An assessment example where a 4-student-training is conducted is given in visual form by means of Matlab.