Food production is an important aspect affecting greenhouse gas emissions. As carbon emission intensity coefficient values of different food types vary greatly, the improvement in economic development levels and living standards and the resulting rapid change of resident food consumption in China have important impacts on the ecological environment and carbon emissions in the country. This study examined the trend of change of urban and rural resident food consumptions in China from 1997 to 2016, and analyzed its spatial-temporal agglomeration characteristics and driving factors with the Lorentz curve and Gini coefficient. The main conclusions are as follows: The total quantity and per capita share of Chinese resident diet carbon emissions were increasing gradually, but they still belong to the low-carbon consumption category globally. Structurally the various kinds of resident diet carbon emissions in China showed different trends of change, and animal food carbon emission increment rate was faster than plant food reduction rate. Residents' diet carbon emissions showed a significant urban-rural division in quantity, structure, and trend, but the gap is narrowing in recent years. Per capita emissions showed significant differences among the ethnic groups and regions, and the high value regions of food consumption carbon emission density were mainly distributed in the municipalities and economically developed coastal areas. Structural optimization was the main direct driving factor of the national food consumption carbon emission change in China, while population was the main factor of regional differentiation.