Abstract: To improve the intelligent perception level of bathing environment comfort and address the limitations of traditional single-modal monitoring. By using multimodal fusion technology, multi-source data including temperature, humidity, wind sensation, air quality, sound, and light are integrated to construct an intelligent perception model based on Multilayer Perceptron (MLP). The model combines feature-level and decision-level fusion algorithms, and analyzes influencing factors such as data quality and model complexity. The model achieves accurate multi-dimensional evaluation of bathing environment comfort, providing data support for intelligent control of bathing appliances. Multimodal fusion technology effectively enhances perception accuracy. The research results lay a theoretical and technical foundation for optimizing the functions of bathing appliances and user experience. In the future, it is feasible to further expand data types and improve algorithm efficiency.

Key words: Multimodal fusion, Intelligent perception, Bathing environment, Comfort evaluation, Bathing appliances