As one of the key parts of rotating machinery, the accurate prediction of its remaining useful life(RUL) plays an important role in the normal production and personal maintenance safety of workers. Due to the complex and Changeable Working Environment of rolling bearings, there are less reference samples in the same working condition and more in the different working condition, which is characterized by unbalance, incompleteness, no label and noise interference, the difficulty of RUL prediction for rolling bearings is increased. In recent years, with the continuous appearance of machine learning and deep learning technology, a large number of intelligent bearing life prediction methods based on neural network have been proposed, the existing forecasting methods are analyzed and summarized, and the future development trend is prospected.