The invention relates to a real-time fault detection method and system of a processing machine tool main shaft bearing. The method includes the following steps that: steps S1, processing part data before and after bearing replacement are acquired and are adopted as a sample set, and at the same time, the normal range of bearing operation acceleration is extracted; step S2, distribution curves of the processing par data in the sample set are fitted, and changed parameters before and after bearing replacement are determined according to the curves, and a part detection risk interval is obtained; step S3, acceleration signals of each shaft of a machine tool and the processing part data are acquired in real time; step S4, if the processing par data reach the part detection risk interval, step S5 is executed; step S5: if the characteristic value of the acceleration signals exceed the normal range of the bearing operation acceleration, step S6 is executed, otherwise, the method skips to the step S3; and step S6, a bearing replacement alarm is emitted. Compared with the prior art, the real-time fault detection method and system of the processing machine tool main shaft bearing of the invention have the advantages of real-time automatic detection, early warning, low misjudgment rate, high equipment utilization rate and the like.