In keyword-driven testing (KDT), a test case depends on a number of keywords, and a keyword depends on some other keywords or GUI components (widgets). Such dependency is also known as the coupling between test cases, keywords, and components. This paper proposes a coupling measure called unweighted coupling and a visualization tool called KTV (Keyword-driven Test script Visualizer) with which the structure of KDT scripts can be quickly grasped. A case study is conducted to assess the maintainability of KDT scripts with various degrees of couplings. The results indicate that, when maintaining KDT scripts, a low-coupling script required, on average, less changes than a high-coupling one. In addition, the estimated change impact offered by KTV was highly correlated with the actual maintenance cost. Thus, in general, the low-coupling principle holds for KDT scripts and can be important to a tester who needs to constantly develop/maintain KDT scripts.