Hydraulic impact hammers are one of the mechanical excavators that can be economically used in tunneling projects under favorable geologic conditions. However, there is relatively less published material in the literature directed to their performance prediction in terms of rock properties. In tunnel drivage projects, there is often the need for accurate means of performance prediction of related mechanical excavators. A poor prediction of machine performance can lead to very costly contractual claims. In this study, the application of a relatively new soft computing method for data analysis called adaptive neuro-fuzzy inference system (ANTIS) to predict net breaking rate of an impact hammer is demonstrated. The prediction capabilities offered by ANFIS were shown by using field data of a metro tunnel drivage project which appear in the published literature. For this purpose, an ANFIS-based prediction model was constructed and the obtained results were then compared to those of regression-based prediction, in terms of various statistical performance indexes. The results suggest that the proposed ANFIS-based prediction method outperforms the classical regression-based prediction method, and thus can be used to produce a more accurate and reliable estimate of impact hammer performance from Schmidt hammer rebound values and rock quality designation values obtained from the field.