SUMMARY Nowadays, the requirement on diminishing operating and maintenance costs is highly concerned for utilities in competitive electricity market. Power transformer is first focused due to its acquisition costs and failure consequences (FCs). Traditional preventive maintenance for power transformer is generally applied. However, it is very costly and does not take into account actual transformer condition. Hence, risk-based maintenance is introduced in this paper to facilitate the maintenance works of power transformers. The conditions of power transformers are evaluated from electrical test, insulating oil test and visual inspection by using the ranking and weighting techniques. In addition, the Analytic Hierarchy Process (AHP) is proposed to determine the important weighting factor of each transformer component. The three main criteria of the AHP are maintenance difficulty, FC and failure history. Moreover, the Weibull distribution techniques are applied to analyze failure causes. The computerized web-application program is developed and implemented for practical use in a utility in Thailand. The power transformers installed in 115 kV and 230 kV transmission systems in Thailand are assessed due to availability and quality of data. From statistical analysis of failure records, the conditions of power transformer can be divided into three different zones as risk, moderate and healthy zone. The defective components in risk zone with higher proportion of failure should be carefully focused. The components and overall condition of the transformers are perceived from the analysis of test results. Subsequently, the risk-based maintenance is properly planned according to the actual condition. Therefore, maintenance of power transformer fleet can be effectively managed resulting in higher availability and reliability, lower risk of failure and lower cost of maintenance. Copyright © 2013 John Wiley & Sons, Ltd.