In view of the inability of the condition monitoring technique of the hydroelectric generating sets to effectively obtain health condition information in real time， an online monitoring approach to the health condition of hydroelectric generating sets based on macro tests is proposed according to the model for the relationship between system composition and behavior and with the help of thought used to grasp the health condition of hydroelectric generating sets by experts through artificial tests. The operating conditions of hydroelectric generating sets in normal running are regarded as tests， which are spoken of as macro tests along with artificial tests. On the basis of complete state monitoring for hydroelectric generating sets， the identification of macro tests， the record of test data， and the calculation of performance indices of macro tests are automatically completed in real time. Through the detection of operating state abnormalities， degraded performance or detection ultra vires， online monitoring for the health condition of hydroelectric generating sets can be realized. This approach has been successfully applied in Gezhouba Hydropower Station， whose practicality and effectiveness are proved by actual operations.