Abstarct: Drilling operation is a difficult and expensive process, which can be pointed out to problems and difficulties such as well blowouts, drilling fluid loss and stuck pipe. and the purpose of this research is to increase the accuracy in the estimation of drilling fluid loss. Also, drilling fluid loss causes an increase in costs, rig waiting time and formation damage. In this thesis, an attempt has been made to obtain the amount of drilling fluid loss during the drilling process of one of the wells in the Azadegan oil field by using multi-laxyer perceptron artificial neural networks. In this thesis, the data of 14 parameters were selected as the input and the drilling mud was selected as the target parameter. The investigated parameters are: well depth, drilling time, pump rate, standpipe pressure, mud weight, solid percent, 10 seconds and 10 minutes gel strength, yield point, plastic viscosity, bit diameter, open hole depth, rate of penetration and drilling fluid loss volume. 70% of the data was used to train the network. 15% of the data were used for evaluation and the remaining 15% were used for testing the extracted network. A feed-forward network called “Lossnet” was designed with specifications according to TRANLIM training function and MSE performance function, including 1 hidden laxyer and 10 neurons for processing input and target information. In the results obtained from the network modeling, it was observed that the correlation coefficient of the network is 0.93 in the training phase and 0.85 in the test phase, which shows the high accuracy of the network in predicting the amount of drilling fluid loss. Also 7.6 % error was seen while matching result with real data.