Abstarct: Pipe sticking is a conventional problem in drilling industry. This matter can cause several problems from time and cost lost, to losing the hole. In this study, tried to making a model to predict pipe sticking. For this task, drilling and fluid parameters of several wells of a giant field have used. 126 case of 973 data sets, that each one is included of drilling and fluid parameters, are related to stuck pipe and others are related to non-stuck pipe observations. In this thesis, at first classification algorithms such as K nearest neighbor and support vector machines, have used to classification of data. KNN 93.6 percent and SVM 97 percent were successful to this task. Then multi laxyer perceptron neural network and neural network time series have used to predict pipe sticking. For first time ever, time series as a dynamic method has used to predict of pipe sticking. Prediction accuracy was 94.9 percent for MLP neural network and 94.2 percent for time series. Being informed to pipe sticking probability could help to prevent it by changing effective parameters.