Life cycle assessment (LCA) is widely used to reduce a building's environmental impacts in the design phase. Buildings consume a lot of electricity, and heat pumps are often proposed as a way to reduce greenhouse gases emissions. The electricity production mix is therefore an important aspect in building's LCA. Many studies use a static national average mix, ignoring its variations in space and time. This might be questioned in a large country undergoing energy transition such as China. A comprehensive study on this topic for China is lacking, therefore this article aims at filling this gap by investigating how the variations of the energy mix at spatial (five regions in five climate zones) and temporal (four future energy mix scenarios) scales influence the LCA results. A model is proposed to evaluate local future energy mixes. The life cycle inventory (LCI) database was contextualised considering different local energy mixes. Environmental impacts calculated using the local energy mixes and the national average energy mix were compared in the static approach and the dynamic approach (future scenarios are considered) for a residential building. The results indicated that using a national average mix instead of a local mix in the static approach brought non-negligible differences for most provinces, e.g. the overestimation of global warming potential (GWP) reached 500 % in Yunnan. Similarly, differences between the static and dynamic approaches are large for most environmental impact indicators, e.g. the difference in GWP could reach around 900 % in Guangdong. The differences highly depend on the prospective future scenarios and showed regional features. This paper highlights the importance of the choice of energy mix in buildings' LCA in China regarding both spatial and temporal scales, which is beneficial for more reasonable decisions.