Reliability test is focusing to detect the unexpected reliability failure and solve them for the high quality of products. The test data should be used to assess and project the current level of interesting product reliability and so it is very important to have the accurately assessing methodology with test data. There are two type of trend for test data as constant and changing one during testing and this paper shows the difference in the assessing results of these two cases. There is less information how to define the existence of reliability growth rate changing and calculate the parameters of the reliability growth models to make an accurate assessment with such condition, so i established the process and mathematical model to calculate the parameters at such condition to make reliability growth curve with high Goodness of Fit. I validated the new method with the data made from Monte Carlo Simulation and case from Demko (1993). Even the assessed result with the new methodology may be different with the case by case because of very diversity in test condition and testing product quality, but the process and method founded in this research can be applied to any case using Duane and AMSAA model for their test data assessment. I also present the evaluation method to see the effectiveness with new one which is a conventional knowledge and not popular to use, so it is possible to compare the results with the newly presented and conventional method for better business decision.