While biofeedback therapy based on occlusal force has demonstrated safety, effectiveness, and minimal adverse effects in bruxism treatment, the underlying mechanism and fundamental technologies remain elusive. This study aimed to implement a biosensor device into a conventional bite night guard to detect bruxism. A layering process (sandwich technique) was utilized to integrate stress sensors, which were then inserted into an acrylic occlusal stabilization splint. The core components of the sensor system consisting of a server terminal, a core control module, and a "pressure sensor data acquisition unit". The occlusal force data analysis and parameter setup employed a machine learning technique. A prototype sensor platform was meticulously developed to assess each aspect of the smart splint comprehensively. Experimental outcomes substantiated the viability of the intended approach for teeth-grinding therapy and unveiled appropriate parameter metrics for the sensor system