In this paper we present a theoretical analysis of the relative error exy in measurements of the mean wind velocity with a cw Doppler scanning lidar. Our analysis takes into account spatial and temporal averaging of the wind velocity fluctuations and the dependence of this error on the turbulent state of the atmospheric boundary layer during the measurements. Our theoretical results have been verified in the experiments with a Doppler lidar at the Institute of Optoelectronics. The experimental data have shown satisfactory agreement with the theoretical results. It is shown that under conditions of stable atmospheric stratification and at altitudes h>150—200 m the measurements of mean wind velocity become representative (exy ≤ 10%) already after one scan. In the ground atmospheric layer (h = 60 m) under conditions of neutral stratification the error exy < 10—12% can be obtained only after five scans (N = 5).