Abstract: The traditional innovation diffusion models which are mostly based on BASS model can't describe the influence of individual personalities, and can't depict the patterns of diffusion in the case that the user abandons the application during the process. For this, this article introduces discrete choice theory into innovation diffusion field to study the influence of personal characters, and combining with cellular automata technology, it builds a hybrid innovation diffusion model. This model is used to simulate the diffusion of Solar Water Heater(SWH)in China's rural areas and it is found that this model can fit the actual diffusion data very well and explain the impact of individual characters on diffusion process. In contrast with the traditional theory, it is discovered that the diffusion speed doesn't display a simple linear relationship correlated to oral communication. Instead, there is a critical value and after this value, the impact of oral communication on diffusion will increasingly decrease.