Dissolved Gas Analysis (DGA) is one of the most common methods used to diagnosis the fault types in the transformers. The fault types which occur in the power transformer are categorized as partial discharge, low discharge energy, high discharge energy, low thermal fault, medium thermal fault and high thermal fault. These faults can be specified according to the concentrations of each dissolved in the transformer oil. In this paper, the modified technique result depends on the outputs of the three techniques (new approach, conditional probability, Duval). The modified technique was prepared to increase the accuracy of the detection of the transformer fault types. It was constructed based on 386 samples that were collected from the Egyptian Electricity Transmission company (EETC) - Central Chemical Laboratories, and were known their actual faults. The result of the modified DGA technique indicated an increase of the diagnostic accuracy rather than the other DGA techniques.