District heating and cooling (DHC) systems have become increasingly popular in recent years in China and all around the world, due to their high equipment efficiency and the capability of using community-scale renewable energy sources. From the perspective of energy use and urban landscape, district cooling systems have great potential value. However, in China, most of the district cooling systems have not performed at their optimal efficiencies. According to some measurement results, a large amount of existing systems that are not performing at optimal efficiencies are oversized. Oversized systems lead to wasted investment, low operational efficiency and, thus, waste of energy. To this end, the thermal load demands of building users in a district need to be accurately understood. In contrast to a single house or a residential building, a district has tens or hundreds of households with varying thermal demands. Spatial and temporal diversity of various households results in different features of central supply load from individual household. In order to analyze the characteristic differences among systems with different scales (e.g., household air conditioner, central DHC system serving tens of apartments, central DHC system serving hundreds of apartments), we collected the household hourly loads from the smart meters in one residential district in Zhengzhou, China for quantitative analyses, which has a central ground source heat pump system to supply the cooling and heating for the three high-rise buildings. Then, we compared the load analyzed the thermal load diversity among combinations of different number of households based on real data from the variation of peak loads, total thermal consumptions and load distributions, as well as the load volatility in typical day. Finally, we also discussed the impact of load characteristics in DHC systems on district HVAC system sizing.