A new fault location algorithm is proposed in this study for a hybrid transmission line, which is a combination of an overhead line and an underground cable line. Owing to the physical and electrical differences, this combination leads to serious challenges for fault identification and location. In the proposed algorithm, using wavelet transform (WT), some key features are extracted from the faulted voltage and current signals. The extracted features are used as inputs for K-nearest neighbour algorithm (KNN). KNN is an intelligent system which is used here for the classification process. For the internal fault identification and location, a four-layer construction of KNN is used. Using the first layer, the internal fault is identified. The second layer of KNN shows whether the fault occurs in the overhead or cable line. The third and fourth layers show which half of the overhead or cable line is encountered to fault. Then the traveling wave theory is used to calculate the exact fault location. In this paper, the network containing the hybrid line is simulated by EMTP-RV software and then the simulated data samples are transferred into MATLAB software. Feature extraction process and identification process using KNN, are performed in MATLAB. Only 2 milliseconds of the input signal is needed for the proposed algorithm and results show that, for most of the cases, the fault location error is smaller than 0.5 km. This shows the fast, accurate and efficient performance of the proposed algorithm.