China is undergoing massive changes in order to tackle their carbon dioxide emissions problem. Aside from the problem of climate change, it is facing another problem – that of air pollution. Use of dirty fuels by residential households such as biomass and coal cause emissions harmful for the environment, as well as to the health of citizens. In order to reduce this, local governments are tasked with inducing a change in household fuel consumption from dirty fuels (such as biomass and coal) to ones that are cleaner and more efficient, such as natural gas or electricity. The ‘lower-tiered’ cities are often left to their own devices in how to implement suitable policies in order to reach these targets. As a result, the policies that are introduced often ensure targets are reached, but also leave citizens without any fuel alternatives, or forces residents to resort to other unhealthy fuels. This research investigates how lower-tier prefectures can stimulate their residents to shift daily energy fuel demand from traditional fuel use to cleaner and more efficient fuels, while preventing the population from unwillingly reducing their total energy demand. The research was performed through a case-study research in the prefecture city of Jingmen in Hubei province. Energy consumption demand behaviour was investigated using a system dynamics model through mechanisms of citizen reaction to energy price level, convenience of fuel consumption, knowledge about clean and efficient energy use and willingness to making a conscious change in consumption. The research finds that the urban citizens are concerned about the effects of fuel consumption on their personal health, while the rural population shows larger concern for the price of fuel and convenience of consumption. A ban on coal will best reduce emissions but it should be implemented in combination with a supportive policy in order to prevent citizens from unwillingly reducing energy demand. Other policies, such as providing an energy subsidy will be less effective and lead to higher levels of uncertainty.