A hybrid depth neural network model is proposed to predict the drillability to increase the drilling speed and reduce the drilling cost. The drillability of rocks can directly reflect the difficulty of drilling in the formation. The prediction effects of CNN(Convolutional Neural Network)-BiGRU(Bidirectional Gated Recurrent Unit)-attention, CNN-BIGRU, and CNN-GRU-attention models on the evaluation indexes of rock drillability before and after adding rock properties were compared, finally, the CNN-BiGRU-attention model with better effect was selected. The model extracts the features of the data through CNN, then the data is processed by the BIGRU for time series features to establish a long-term dependency on the features. Finally, the Attention mechanism is introduced to focus the data to improve the accuracy of the model prediction. The coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE) were used as the evaluation parameters of the model. The results show that the prediction accuracy of the deep neural network model is improved by 0.06, 0.08, and 0.1 when the formation properties are added as input parameters to the model. The results show that the physical properties of the rock have an important effect on the drillability of the formation. Therefore, taking the rock properties as the input of the model can improve the accuracy of the model, and the CNN-BIGRU-Attention model has a good effect. It is believed that this method can be popularized in areas similar to this formation and is conducive to guiding the block to improve drilling efficiency.