The failure rate of the machine tool depends on its inherent reliability. Although the operating environment influencing factors will not change the failure law, it will advance or delay the cycle of the failure to a certain extent. Therefore, it is very important to consider the environmental impact factors when making machine tool spare parts prediction. By analyzing the operating environment of a certain type of machine tool spindle, the main influence covariates were screened, based on the hypothesis test of the data independent and identical distribution, the PHM regression model was used to establish the reliability model of the cylindrical roller bearing. Based on this, the updated model was used to estimate the point of failure, by referenced to the spare part availability, the establishment of a demand forecast model for spindle spare parts was completed. The calculation results show that it is more accurate to consider the environmental factors for the prediction of spare parts. The research results have theoretical reference value of study for the prediction of irreparable spare parts for machine tools.