In order to improve the automatic operation degree in traditional peanut combined harvester, the online conditions measurement method for key working parts, such as picking roller, sorting screen, clamping shaft, clamping shaft and rotating components, was designed on the basis of 4HBLZ-2 one-row small self-propelled peanut combined harvester. Using LabView, an on-line operation monitoring system of peanut combined harvester was developed, with which the machine control state, harvesting operation mode, engine parameter, harvesting trajectory and working conditions of main parts could be real-timely monitored. The self-diagnosis and fault early warning model was established with the application of multi-sensor information fusion algorithm, which could offer alarm message for drivers under abnormal conditions, such as blocking of picking roller and chain break. Tests result showed that the time of automatic fault diagnosis was less than 2 min and the accuracy of fault detection was up to 90%. The designed system had met the functional requirements and precision needs of peanut combined harvester real-time monitoring in field operation.