A projection image analysis method for the fault diagnosis of rolling bearing based on Intrinsic Mode Function（ IMF） is proposed,the status recognition and the degree of recognition of the rolling bearing fault is realized. The main contents are as follow. Firstly,the Empirical Mode Decomposition method（ EMD）is used to obtain IMF. Secondly,the time-frequency three-dimensional gray-scale projection image of each IMF is established. The texture feature of Gray Level Co-occurrence Matrix（ GLCM） is also introduced to analyze the three-dimensional projection image. Finally,the feature dimension is compressed and combined based on the principal component analysis. The Support Vector Machine（ SVM） is used to realize the fault diagnosis of rolling bearing. Both fault diagnosis and degree identification of rolling bearing is realized through a perspective of image characteristics. A new method based on projection image of characteristics calculation in vibration signal is given in this study.