Fuzzy inference systems have widely been applied to model complex, high dimension non-linear real-life problems. The common strategy for designing type-1 and type-2 fuzzy models is to first cluster the data, followed by optimising the fuzzy set parameters and the rule-base. Besides being computational intensive, these design methods also treat the fuzzy system as a black-box model. The input-output relationship is not fully understood. Therefore, in this paper, the analytical structure of a simple two-input-one-output interval type-2 fuzzy logic system is used to analyse the effect on the fitting ability of the fuzzy model when one of the parameters of the antecedent fuzzy set is varied. Results showed that it is possible to predict whether the predicted output increases or decreases when a particular FOU parameter is varied. Numerical examples are given to demonstrate that the modelling accuracy for a range of input conditions can be improved by varying the parameters of the antecedent fuzzy set. Improvement is supported by a reduction in RMSE. The work reported in this article is the first step towards a better understanding of the effect of varying individual FOU size in the input domains.