In view of the complexity and uncertainty of evaluation indicators for power transformer condition assessment, the theory of fuzzy set pair analysis was adopted.Combining actual situation of power transformer condition evaluation area, following the indicator select principle which includes scientificity, systematicness, representative and so on, a power transformer condition evaluation index system and evaluation grading standard were constructed, and the main features of the complex decision-making process and the identical discrepancy-contrary (IDC) system were analyzed to establish the connection degree formula.The connection degree of evaluated factors was calculated and a new set pair analysis(SPA) model was built, and the weights of evaluated factors in the model were obtained from fuzzy analytic hierarchy process.A transformer was taken as an experimental site through repeating trials in comparison with the other related methods.The results indicate that fuzzy set pair analysis model conforms power transformer condition evaluation, and it is simple and convenient. It has fuzzy synthetic properties, and can utilize adequate information to gain the reliable evaluation results through mathematical calculation, thus can provide a preliminary evaluation of power transformer condition.