Sugeno Integral is based on Fuzzy Integral Inference and widely used in applications such as decision making and computational intelligence. When concerned inputs are intervals, directly using Sugeno Integral to respectively aggregate the lower bounds and upper bounds of those intervals has limitations and does not embody fuzzy integral inference. This study analyzes the fuzzy integral inference in interval environment, defines some more suitable orderings on the set of all intervals in [0, 1], i.e., the congruent $\lambda $ -ordering, and then proposes the Interval Sugeno Integral with preference. The novel aggregation technique proposed in this study proves to better embody fuzzy integral inference when performing Sugeno Integral.