Abstract：The rock drillability is an important index for optimizing bit selection and programming drilling parameters. Prediction of rock drillability before drilling is very important for the increase of drilling efficiency and the reduction of drilling cost in deep wells or super-deep wells. Nowadays，at home and abroad，there are few reports about how to predict rock drillability of wildcat well before drilling in new exploration area. For wild well，there are lack of logging data and core data. Therefore，present evaluation methods of rock drillability can not be applied to predict rock drillability of wildcat well before drilling. Based on the selection of similar structure，a method for predicting the rock drillability of wildcat well is proposed. GA-BP (genetic neural network) model is established by use of seismic data，logging data and core data of the similar structure well. On the basis of neural network theory，genetic algorithm is used to optimize neural network. According to seismic data，rock drillability of wildcat well will be predicted before drilling by use of this GA-BP model. Rock drillability for Xinjiang well MX1 is predicted before drilling. Compared with evaluation results of logging data，average relative error of the prediction result is 9.8%. The field application result testifies that this method is feasible and has a high accuracy.
