Operation conditions usually change when rotating machinery works. The changing operational conditions and machine fault can make the mechanical vibration characteristics change and cause diagnosis errors, so a new method for the health monitoring of bearings under changing operational conditions was proposed. In this method, the RVMs (Relevance Vector Machines) were used for obtaining the continuous function relationships between the adaptive parameters of the threshold model and the characteristic statistics of vibration features. Based on the characteristic statistic in different operation conditions, the adaptive threshold model was constructed. This method was used for bearings health monitoring at different revolving speed. Monitoring results show that this method is effective only when the rotational speed is higher than a relative small value.