Summary: The reliability estimation of the components in series system is investigated using masked system life data under multiple type-II censoring samples. At first, the sample likelihood function is deduced using the probit analysis method. Then based on the sample likelihood function, maximum likelihood estimations of component parameter and reliability function are obtained. Meanwhile, Bayes estimations of component parameter and reliability function are given under squared error, linear exponential and general entropy loss function, respectively. Finally, a numerical simulation is introduced to study the influence of masking level and censoring factor on the accuracy of the estimations and compare these estimations.