<p>Industrial instrument panels serve as the primary interface for human–machine interaction in safety-critical environments. They are often designed with varying shapes, scale types, and number sizes to accommodate spatial constraints and functional requirements. A common practice is to select these features based on empirical conventions or aesthetic preferences. However, whether such composite configurations preserve perceptual consistency and minimize cognitive load under industrial monitoring conditions remains unclear. This study experimentally evaluated visual search efficiency and cognitive load across composite industrial instrument designs. We focused on three interface features (instrument shape, scale type, and number size) and examined their independent and interaction effects using a three-way ANOVA, complemented by eye-tracking and NASA-TLX assessment. The main findings are as follows. First, shape dominates visual search efficiency: arc-shaped instruments, which follow natural saccadic trajectories, achieve the shortest reaction time and the lowest cognitive load, whereas horizontal instruments perform the worst. Second, scale type and shape interact non-additively: progress bars reduce reaction time via intuitive processing, but arc-shaped instruments with pointer scales yield the highest accuracy. Third, among 10-pt, 14-pt, and 18-pt numerals, 14-pt optimizes performance for arc-shaped layouts while preserving visual aesthetics. These findings refine the Gestalt continuity principle and dual-process theory in the context of industrial human–machine interaction, and provide an evidence-based framework for industrial instrument design.</p>