The present study provides a feasible method to evaluate creep properties for a 9%Cr-Mo-Co-B power plant steel by comparing two sets of data obtained from small punch tests and conventional uniaxial creep tests. The method includes three steps: firstly, conduct a series of small punch tests and conventional creep tests in different load and temperature conditions; secondly, convert the load and central deflection data obtained from the small punch test to stress and strain data; thirdly, determinate the best fit correlation factor by comparing the two sets of data in selected creep models. It is found that two sets of data show a similar trend in stress–rupture time relation, stress–minimum strain rate relation and LMP–stress relation. The correlation factor, ksp, can effectively bridge the gap between the load in small punch test and the stress in conventional creep test. For a high-Cr martensitic heat-resistant steel named as CB2, the ksp value 1.4 can make a good prediction for rupture time, while for minimum creep rate and the Larson–Miller parameter, the ksp value 1.4 will lead a conservative prediction in the low-stress range.