In view of the highly randomness and uncertainty in the working condition of crane, take bridge crane as the research object. Firstly, based on the technology of the internet of things, the load capacity and the number of work cycles would be recorded, and the fatigue stress spectrum would be formed. Secondly, based on the Miners fatigue damage accumulation theory and the rain-flow algorithm, the equivalent stress amplitude would be obtained. Thirdly, curve regression model has been used to characterizing the relationship between the crack propagation and equivalent stress amplitude, and predicting the current crack size. Lastly, taking the predicted value of the crack size into the fracture mechanics formula, and estimated the remaining fatigue life of the bridge crane. The example demonstrated that, it is simple and practical to apply the techniques of the internet of things and the regression forecasting to the data collection and crack size prediction; it not only be able to estimate the remaining fatigue life quickly and accurately, but also be able to overcome the drawback of requiring the initial crack size in the fracture mechanics.