In terms of relative normalized gain array (RNGA) criterion, this paper presents an interaction analysis and loop pairing method for multi-input-multi-output (MIMO) processes described by Type-2 Takagi-Sugeno (T-S) fuzzy models. Type2 fuzzy system offers a significant improvement on traditional (Type-1) fuzzy systems to handle the uncertainties. For each individual loop in the MIMO process, the steady-state gain and normalized integrated error can be derived from its Type-2 T-S fuzzy model through simple formulae, and the pairing results can then be obtained according to the rules of RNGA criterion. This method can be applied to measure the interactions and determine the control structure for multivariable control system design when the mathematical functions are unavailable and a large number of uncertainties present. The simulation proves that the proposed method can provide more accurate results than the existing RNGA paring method based on Type-1 fuzzy model under the influence of uncertainties.