In this study, we developed a sensor-integrated evaluation framework for wayfinding visual communication in intangible cultural heritage exhibitions and validated it through a case study of the Shanghai Jinshan peasant painting exhibition. The objective of this study was to establish a quantitative and reproducible sensing-based framework for evaluating and optimizing wayfinding visual communication systems through the integration of eye tracking, context-aware sensing, task-performance analysis, and subjective evaluation. The proposed framework combined subjective ratings, task-performance indicators, eye-tracking metrics, and context-aware sensing variables to enable a comprehensive and quantitative assessment of visitor behavior, visual attention, and environmental influence. A two-stage experimental design was adopted. In Stage 1, an optimized wayfinding system was developed through design-principle review, expert consultation, and iterative pilot refinement. In Stage 2, a quasi-experiment was conducted with 60 participants, who were assigned to the original and optimized wayfinding groups (n = 30 each) and asked to complete five exhibition-guidance tasks. Multisource data collection included subjective scales, task completion time, task accuracy, time to first fixation, area-of-interest (AOI) attention ratio, and scan path concentration, as well as environmental sensing variables such as crowd density, illuminance, and viewing distance. The results showed that the optimized version significantly improved wayfinding clarity, exhibition understanding, and overall evaluation, with large effect sizes. It also significantly reduced completion time and improved task accuracy across all five tasks. At the eye-tracking level, the optimized version shortened the time to first fixation and increased the AOI attention ratio and scan path concentration, indicating more efficient visual guidance toward key information areas. Correlation analysis further revealed that better eye-tracking performance and task performance were associated with higher subjective evaluations. Regression analysis identified the wayfinding version as the most stable predictor of visual attention efficiency, task performance, and exhibition understanding/overall experience. These findings demonstrate that the proposed framework provides a measurable and reproducible approach for evaluating exhibition wayfinding systems and offers empirical evidence supporting the sensor-based optimization of visual communication in cultural exhibition environments.