Purpose This study aims to introduce a new modified method for estimating steam turbine high pressure (HP)-intermediate pressure (IP) leakage flow based on the experimental data extracted from a 250 MW re-heat steam turbine. Design/methodology/approach Effects of measurement uncertainties on the test results are investigated and key parameters are specified via a new modified method to diminish the test uncertainties. The recommended method is based on a constant IP turbine pressure ratio at the same loads. Considering this assumption, it was found that the turbine pressure ratio can be achieved in the second and the third tests with a much longer duration. Findings The results showed that the cross-over temperature is a major parameter in the leakage flow estimation tests, whereas hot reheat and cross-over pressures are the next priorities. It was also observed that as the cross-over temperature varies by 1°C, the estimated leakage flow error significantly differs by up to 72.6 per cent. Originality/value It is concluded that the present modified HP-IP leakage flow estimation method seems to be more accurate in comparison with previously proposed methods in literature.