In order to enhance the competitiveness of the domestic hydraulic support. It is imperative to carry out reliability engineering. The reliability evaluation of the hydraulic support in service can effectively provide reference for the development and improvement of the product. It can also provide a reference for development of new type of reliability test. Based on support vector machine has great advantages in small sample data processing,Putting forward using support vector regression machine estimate weibull distribution parameter of hydraulic support failure time,when the failure date is small sample. The accuracy of the results obtained by different empirical formulas is compared. The reliability of the model is evaluated and predicted by the highest accuracy. Research shows that: support vector machine can be used to evaluate the reliability of hydraulic support under small sample maintenance data. Shape parameter of some commonly used hydraulic support reliability Weibull distribution is3. 0157,the scale parameter is 61728.