We studied random error of horizontal wind velocity estimated by the spaced antenna (SA) technique by means of computer simulations with one spatial dimension. Complex time series of data were generated for the no-noise case and were processed through the full correlation analysis (FCA) to estimate horizontal wind velocities. The simulation parameters were chosen to simulate typical observations with the middle and upper atmosphere (MU) radar. For simplicity we assumed isotropic scattering in the simulations. Comparing performances of the SA and Doppler techniques, we found that the estimation error of the SA winds is larger than that of the Doppler winds in most cases. For large wind velocities, however, the SA technique showed better performance than the Doppler technique. By using the MU radar, we conducted simultaneous multibeam/interferometry observations to evaluate the error of wind estimation from real data. Performance of the SA winds obtained from the observations was seen to be similar to that obtained from the computer simulations.