Summary: The excessive test time or a large sample size is required to demonstrate the reliability of long life time highly reliable products through conventional zero-failure testing. In order to reduce the test time, a method for accelerated degradation zero-failure reliability demonstration testing based on performance degradation data under binomial distributions is proposed. First, the degradation and its variance are estimated at the specified time of a test unit from an accelerated degradation model and measured data. Then optimum test plans are derived by minimizing the total test cost, which satisfy the constraints on the consumer’s risk and available sample size. The decision rules for terminating the test of a unit are also presented. Finally, the proposed method is validated to be effective on reducing the test time with a real example. A sensitivity analysis indicates that the pre-estimation of degradation during the specified time period has a great effect on the optimum test plans.