Traditional life demonstration test is conducted under normal conditions and fails to use degradation observations，thus requires large amount of time and money. We can employ degradation data to predict whether a unit will pass the test by the end of test，and the test can be terminated early when this prediction does not cost unacceptable risk. Also higher stress can accelerate the degradation of the unit to cut down the test time. Thus it is necessary to propose a method that take advantage of both degradation data and accelerated test to reduce the test time and cost，which is done by this article. In particular，the article propose an approach to analyses and calculate the additional risks caused by early termination of the test and applying higher stress based on degradation model，then the minimum sample size of different test time can be picked. The article also use optimal plan to give expected test time and give deciding rules of terminating the test for each unit. At last，the method is illustrated by an example， showing that the method is efficient in reducing test time.