Novel extraction methods such as microwave-assisted extraction are fast and effective methods for extraction of nutraceutical from plant tissues. Correct selection of extraction method caused to extract the most concentration of phenolic compounds from certain material in compared to traditional methods. In this study, effect of ethanol concentration (10, 20, 30, 40, 50, 60, 70, 80, 90 and 100%) and extraction time (0.5, 1, 2, 5, 7, 9, 11 and 13 min) evaluated on extraction of phenolic compounds from eucalyptus leaves by microwave and neural network used to estimate of process extraction. Results showed that 50% ethanol had maximum extraction content. Multi-layer perceptron (MLP) with Levenberg-Marquardt alghoritm used to education of patterns. Variables of artificial neural network included neuron number in hidden layer (4-30) and activation function type (TANSIG and LOGSIG) in hidden layer. In order to develop ANN's models, ethanol concentration and extraction time are used as input vectors and phenolic compounds content of eucalyptus leaves as the output. The results showed that MLP with TANSIG activation function and 16 neuron in hidden layer can estimate phenolic compounds with R2 of 0.9632 and MSE of 0.00069, in compared with other networks.