Interval is an alternative to probability distribution in quantifying uncertainty for sensitivity analysis (SA) when there is a lack of data to fit a distribution with good confidence. It only requires the information of lower and upper bounds. Analytical relations among design parameters, design variables, and target performances under uncertainty can be modeled as interval-valued constraints. By incorporating logic quantifiers, quantified constraint satisfaction problems (QCSPs) can integrate semantics and engineering intent in mathematical relations for engineering design. In this paper, a global sensitivity analysis (GSA) method is developed for feasible design space searching problems that are formulated as QCSPs, where the effects of value variations and quantifier changes for design parameters on target performances are analyzed based on several proposed metrics, including the indeterminacy of target performances, information gain of parameter variations, and infeasibility of constraints. Three examples are used to demonstrate the proposed approach.