Low stress interrupted creep test, as an interim compromise, can provide essential data for creep deformation design. However, there are no clear guidelines on the characterization of the terminating time for interrupted low-stress creep test. To obtain a suitable terminating time in terms of economy and effectiveness, long-term creep strain data of 9%Cr steels are collected from literatures and their creep deformation characterization is analyzed. First, the variations of normalized time and strain of each creep stage with the stress level are discussed. Then, the effect of the terminating time on final fitted results of Norton–Bailey equation is estimated. Third, the relationship between demarcation points at different creep stages and minimum/steady-state creep rate is analyzed. The results indicate that when the creep rupture life is considered as an important factor for creep design, the tertiary creep stage is of greatest significance due to the largest life fraction and creep strain fraction at low stress level. However, the primary and secondary creep stages are of great significance for design due to their larger contribution to 1% limited creep strain. And the long-term secondary creep data could be extrapolated by combining the primary creep strain data obtained from interrupted creep tests with the time to onset of tertiary creep derived from a similar Monkman–Grant relationship.