针对滚动轴承状态监控中最优共振频带难确定的问题,提出一种新的寻优方法和目标函数,可快速准确地定位共振频带,提取状态信息.通过两步网格搜索法,以包络稀疏性为目标函数,对Gabor滤波器参数寻优;然后对振动信号进行滤波并得到信号包络;最后运用包络自相关谱抑制噪声,突出运行状态信息.用仿真信号和实际信号对该方法进行了验证,结果表明,该方法能准确判明轴承运行状态. Aiming at the difficulty of determining the optimal resonance frequency band in the state monitoring of rolling bearings, a new optimization method and objective function are proposed to locate the resonance frequency band quickly and accurately and to extract the state information.By the two-step grid search method, As the objective function, the parameters of the Gabor filter are optimized, then the vibration signal is filtered and the signal envelope is obtained, and finally the envelope auto-correlation spectrum is used to suppress the noise and highlight the operating state information. The verification shows that the method can accurately determine the bearing operating status.