H pull-out test is widely employed as a primary method for estimating the adhesion value of the interface in the rubber/cord composites. However, the current international standards only require to record the pull-out force in rubber/cord H pull-out tests, and emphasize that results obtained from samples of different sizes are not comparable. In this paper, we investigate the comparability of H pull-out results from rubber/cord samples of varying samples and propose an evaluation method to normalize these results. Rubber/cord H pull-out tests of samples with the sizes of 5 mm, 6.4 mm, and 10 mm are undertaken respectively (the results of the specimens with dimensions of 7.5 mm and 8.5 mm are provided as supplementary information). Based on the test results, the adhesion values are analyzed by the model of cohesive elements in finite element method, and the cohesive parameters derived from these models are considered more accurate indicators of adhesion value than the pull-out force. Therefore, a more reasonable comparison of adhesion values becomes feasible. The cohesive parameters are initially determined using the 6.4 mm model and then applied to the other sample models. Good consistency can be seen between the simulations and the test results even though the sample sizes are changing, demonstrating that cohesive parameters remain unaffected by sample dimensions. Based on this, we develop a normalization method that enables the comparability of rubber/cord H pull-out results from samples of different sizes. The cohesive parameters provide a more accurate representation of the adhesion value, and the normalization method we propose serves as an extension of current experimental methods outlined in international standard.