Acoustic emission (AE) signals were acquired from machining experiments to develop deep learning-based neural network models for predicting tool flank wear on Inconel 617 alloy. The milling experiments were conducted with different combinations of process parameters until the flank wear of the tool reached the predetermined failure limit. The raw data from the sensor signal was filtered to remove anomalies present in the dataset. The Local Outlier Factor (LOF) algorithm was applied to evaluate the anomaly detection performance on the sensor dataset and was found to be the most effective. Statistical features that correlated with tool flank wear were extracted from the filtered data using feature engineering. A deep neural network was developed and its performance was evaluated using a new set of milling conditions. The model’s accuracy was evaluated by comparing the predicted tool wear to experimental results. The results showed that the acoustic emission data can explain 96% of the total variation in the predicted tool wear values.