This work intends to analyze the inaccuracy which appears in the predictions of the conventional gas permeation testing (GPT) method. A theoretical combined with an experimental analysis were presented to study the capability of the conventional GPT method with reference to the membrane structural parameters. Although, the conventional GPT method is relatively easy and fast to be implemented, it normally prone to some inaccuracy caused by inadequate experimental and theoretical techniques. Therefore, it suffers from several conceptual drawbacks. In comparison to the Wakao et al. model, it can be concluded that the modified r(p) and (epsilon/l(p)) were different from the conventional one by a factor of 3 pi/16, and (16/3 pi)(2) respectively. In addition, depending on which part of the J versus P is selected, different r(p) and (epsilon/l(p)) will be resulted in. Finally, at the regions which the slope is negative or zero, the conventional GPT method is not capable to establish any prediction. (C) 2013 Elsevier B.V. All rights reserved.