This study examines farmers' intentions towards pro-environmental behavior in a famous tourist village in China called Guanshan, whose ecological environment is polluted. By adopting the empirically validated norm activation model (NAM) of Schwartz and merging it with Vroom's expectancy theory, the current research aims to develop a refined framework for understanding the formation of and predicting changes in pro-environmental intention. Field surveys were conducted in Guanshan, which resulted in sample data consisting of 275 valid responses collected by the research team. We develop a refined model, including six latent variables and 24 observational items. The structural equation modeling results show that the final model enjoys a better predictive accuracy of pro-environmental intention than does the original NAM. The study also discovers that the motivational force of expectancy theory significantly influences pro-environmental intention, whose motivational force comes from the impact of valence and expectancy. The pro-environmental intentions of farmers are mainly affected by the environmental personal norm and by a certain degree of personal expectancy. The improvement of farmers' pro-environmental intention needs be promoted in two approaches: the cultivation of personal environmental protection norms and the guidance of producing a desired expectation for pro-environmental intention.