The unconfined compressive strength (UCS) is one of the most important strength characteristics of rock which is often predicted by indirect methods due to difficulties in standard sample preparation and to save substantial effort in test procedure. This study presents the prediction of UCS from index properties of gneissic rocks using simple regression analysis, multiple regression analysis and artificial neural networks (ANNs) method. UCS and Index properties such as dry density, saturated density, grain density, porosity, primary and secondary wave velocity were determined from the laboratory tests for nine variants of gneissic rocks. Regression models were developed to establish empirical relationships between these index properties and UCS. Concurrently, ANNs were employed to capture complex, non-linear interactions among the variables for more accurate UCS predictions. Outcome from the regression analysis and ANNs model showed strong relationships between UCS and the index properties in Gneissic rocks. The coefficient of determination (R2) and root mean square error (RMSE) of the predicted UCS of the test datasets by multi regression analysis and ANNs model were 0.92, 5.2 MPa and 0.97, 2.3 MPa, respectively. By comparing the results, it is observed that ANNs performed better than regression models in terms of accuracy. As a result, the prediction equations and ANNs model of this research can be applied to geotechnical applications in regions with gneissic rock formations for predicting UCS.