Gully erosion is a type of water erosion that causes significant sedimentation in watersheds and damages in agricultural lands, rangelands, and infrastructures. This study was conducted to determine the potential of gully erosion by artificial neural network. The Levenberg-Marquardt (LM) algorithm and Multi-Layer Perceptron were used employing soil, geology, land use, distance to fault, slope, aspect, distance from roads, distance from drainage, and elevation data as its variables. Results showed that the structure of 1-13-9 with sigmoid activation function in the hidden layer is more suitable for gully erosion potential assessment. Zonation of gully erosion revealed that the watershed area was divided into different classes of different extent, including 70. 26% in very low, 1. 71% in low, 2. 45% in medium, 2. 65% in high, and 22. 93% in very high potential class. Furthermore, results indicated that slope less than 10%, 50 m distance from the stream, rangeland area, and lithological units of EM and M2 had the greatest impact on the occurrence of gully erosion.