As people’s demand for living environment comfort and personalization continues to increase, traditional indoor space layout methods have problems such as slow response, single style, and difficulty in dynamic adjustment in terms of design detail control and user personality expression. To this end, this paper introduces the integration of artificial intelligence and Internet of Things technology to construct a personalized layout design model for indoor space based on generative adversarial network (GAN). Personalized data such as user behavior path, space usage frequency, and style preference are collected through smart terminals and sensing devices, and encoded into controllable conditions and input into the GAN model to achieve intelligent division and layout generation of spatial functional areas. The system strengthens the control of functional rationality, aesthetic consistency and behavioral flow through a multi-dimensional loss function, and introduces a user interaction feedback mechanism to support dynamic optimization and iteration of layout results. Experimental results show that the generator loss of the proposed method is significantly lower than that of the basic GAN model at each training stage, and the decline rate is faster and the convergence is smoother. In the initial stage (the 10th round), the G_Loss gap between the two is large, and the proposed method is 2.34, which significantly enhances the personalization and intelligence level of the design, and provides an efficient and controllable new path for smart home design.