Abstract The machining of a cone frustum as specified in National Aerospace Standard (NAS) 979 is widely accepted as a final performance test for five-axis machining centers. Although it gives a good demonstration of the machine’s overall machining performance, it is generally difficult to separately identify each error source in the machine from the measured error profile of the finished workpiece. This paper proposes a set of machining tests for a five-axis machine tool to identify its kinematic errors, one of its most fundamental error sources. In each machining pattern, a simple straight side cutting using a straight end mill is performed. The relationship between geometric errors of the finished workpiece and the machine’s kinematic errors is formulated based on the kinematic model of a five-axis machine. The identification of kinematic errors from geometric errors of finished workpieces is experimentally demonstrated on a commercial five-axis machining center, and the estimates are compared to those estimated based on ball bar measurements.