Transformers are one of the most expensive components of power systems and their sudden departure from the circuit will cause great damage. Therefore it is essential to evaluate and estimate their remaining life. Life of a transformer is estimated at the design stage, but due to different climatic and operating conditions, it is different with its real life. Transformer’s life is proportional to its insulation life; in transformers insulations are divided into two types of solid and liquid insulations. Transformer insulating oil can be refined and replaced if it is old or degraded, but cellulose insulation normally can not be replaced and repaired, For this reason, the life of transformers are equal to the life of their cellulosic insulation life. The cellulosic insulations in the transformers are also divided into two types: the preoards and paper insulations. The papers have less strength than preoards, so their life is shorter and is equal to transformer’s life. The best way for estimating the remnant life of these paper insulations is to sample the insulations and to have degree of polymerization experiment on them. But it is required to turn off and open the operating transformers that usually it is not possible. For this reason, many researches have been done to find the aging and cellulose degrading effects in the transformer oil. They have found furfural as the best material for describing the aging of cellulose papers. Then, some methods for estimating the remaining life of the cellulosic insulations based on the oil experiments in acceleration aging have proposed but these methods, because of being just experimental, are not accurate and do not have a suitable answer for real transformers. For these reasons, in this thesis for estimating the transformer remnant life, a combination of methods is proposed. The aim of this thesis is to develop an algorithm for estimating the remnant life of power transformers, based on the field experiments on the transformers. Toward this goal, four methods are used. These methods are as following: structural artificial neural network, life estimation from the top oil temperature, using the transformer health index and life estimation from the furfural concentration dissolved in oil. These four methods are needed to performvarious experiments and measurements.Theseexperiments are chosen byassessing the elements that have effectson the life of transformer insulations and experiments that are useful for measuring these effects. Finally, an algorithm that uses a fuzzy system is designed. This algorithm estimates the remnant life oftransformers, by inserting the four methods results as its inputs. The final results from this algorithm are compared with the other usual methods based on experimental works. In addition, these results are compared with the measured ones from the real transformers. The results of this comparison show that the algorithm can estimate the remnant life of transformers properly, except in the cases that aging is not significant.Furthermore, in this thesis, some samples of oil and paper insulations of some power transformers have been taken and these samples have been investigated in the chemistry labratory. Keywords: Transformer, Estimating Life, Furfural, Degree of Polymerization, Neural Network, Health Index, HotSpot temperature