Compared with the equipment performance degradation prediction under the complete data, the prediction under the missing data is more difficult and more meaningful. However The existing prediction methods of bearing performance degradation do not consider the prediction under missing data. Based on the above problem, a bearing degradation prediction method based on infinite hidden Markov model (iHMM) is proposed under the missing data. In the proposed method, an iHMM prediction model with wavelet entropy as the degradation feature is established to predict the missing data points of rolling bearing sample data and form new complete data. Then the proposed prediction model is used to make single-step predictions on the new complete data. The experiment results show that compared with the real value, the obtained prediction data has a smaller average error. Compare the real value, the predicted value under the complete data, and the predicted value under the new complete data, the prediction data obtained by the iHMM prediction model can also well reflect the degradation trend of rolling bearing. The proposed method can provide a new idea for predicting the degradation trend of rolling bearings under the missing data, therefore the proposed method has important theoretical value and engineering application value.