A method for predicting a traffic situation for an automatically operating vehicle involves constantly recording the surroundings of the vehicle and predicting a traffic situation of the vehicle is predicted for a future point in time using recorded surroundings data and prediction parameters. Upon reaching the future point in time, a current real traffic situation is recorded, the current real traffic situation is compared to the predicted traffic situation, and it is determined in the comparison whether a prediction error was present during the prediction. In the event of a prediction error, the prediction parameters are corrected.