The pretension for a ball screw is a way to improve the position accuracy. Hollow ball screw without a cooling system has the thermal deformation effect due to increase in temperature. It will reduce precision accuracy in machine tool when the ball screw nut preload or ball screw pretension is lost. The purpose of this study is to use vibration signals for the prognostic analysis for the ball screw pretension. Features of different pretension conditions by 0, 5, 10, and 20μm are discriminated by empirical mode decomposition (EMD), fast Fourier transform (FFT), and marginal frequency method. Temperature effects with long-term operation were discussed. This study experimentally extracts the characteristic frequencies for bettering pretension through the vibration signals. This diagnosis results realize the purpose of prognostic effectiveness on knowing the hollow ball screw preload loss based on pretension data and utilizing convenience.