The dynamic compliance (frequency response function - FRF) of a machine tool structure in the cutting zone under a cutting load is one of the major dynamic characteristics that define a machine’s cutting performance. The roundness and surface finish define the quality of the manufactured parts. These characteristics are developed during finishing and semi-finishing cuts. The kinowledge of machine tool dynamic compliance, defined in these steady-state cutting conditions, ensures parts quality and increase in machine tool productivity. The dynamic compliance is usually evaluated in tests, which are performed by means of hammers or vibrators (exciters). During these tests the machine does not cut and the machine components do not move relative to each other. The loads in the machine during cutting are defined by different internal and external sources that are acting in different points of the machine and in different directions. The real spectrum and frequency range of these forces is unknown. Experimental data acquired by different types of tests clearly show the difference in dynamic compliance for the same machine tool during cutting and idling. The machine tool dynamic tests performed by different types of external exciting devices do not take in consideration the real load conditions and interactions of moving components, including the cutting process itself and external sources of vibration. The existing methods of experimental evaluation of machine tool dynamic compliance during steady-state cutting condition require dynamometers to measure the cutting force and a special sensor to measure relative displacement between the cutting tool and workpiece. The FRF that is computed from these measurements represents a dynamic characteristic of the close loop system (machine structure and cutting process) and only under certain conditions may be considered as FRF of machine tool structure itself. The theory of stationary random processes allows defining the cutting conditions, under which the obtained data represent the FRF of machine tool structure, and provide estimations of random and bias errors of this evaluation. The simplified methodology of FRF estimation, based only on measurement of the spindle and tool vibration, is also presented in this paper. This methodology is used on an assembly line to obtain FRF for machine tools performance comparison and quality assurance.