One of the most popular chemical tests for power transformer incipient faults detection is the dissolved gas analysis (DGA). This study presents a new DGA approach for transformer fault interpretation with increased accuracy. The suggested approach is constructed based on the combination of the outputs of the two recently developed DGA techniques; 1) conditional probability and 2) artificial neural network, with two scenarios. These scenarios based on the diagnosis of the same techniques; artificial neural network (ANN) and Conditional probability techniques but with different procedures. This presented approach improved the overall accuracy compared with the existing DGA techniques. The positive feature of the proposed approach is proven by using 240 datasets received from the Egyptian Electricity Transmission Company (EETC) with known faults. The proposed approach concluded that the integration between different DGA techniques with various fault accuracy improves the overall accuracy of fault detection. The suggested fault diagnosis approach, with the two scenarios, exhibited higher accuracy compared to both ANN and conditional probability technique. Also, 42 samples have been used as testing samples for validating the proposed DGA approach, which result in its superiority compared to ANN and conditional probability technique.