Condition assessment is the key technology to realize condition-based operation and maintenance of gas turbine, and it is the focus of current research on intelligent gas turbine. A fuzzy variable weight hierarchical condition assessment method for marine gas turbine is proposed in this paper. Firstly, the hierarchical structure of gas turbine is divided by analytic hierarchy process. The health baseline model of gas turbine thermal parameters based on least square method is established, and the fuzzy condition assessment method of gas turbine thermal parameters is studied based on kernel density estimation, and the health degree model of fuzzy parameters is established. Finally, the optimization algorithm is combined with the variation coefficient method to carry out the weight optimization research of gas turbine parameter level based on the deterioration degree of equipment level, and the gas turbine equipment level condition assessment model is established. Based on the degradation data of gas turbines, a fuzzy variable weight hierarchical condition assessment method for gas turbines is verified. The results show that the method overcomes the traditional weight determination method that relies on expert experience, combines the advantages of fuzzy theory, variable weight optimization method and analytic hierarchy process, and can realize anomaly monitoring and quantitative assessment of the degree of degradation of gas turbines.