Abstract: The cluster conservation of traditional villages has become an important topic in the research of historical and cultural heritage conservation in China. Selecting the Hehuang region as a study case, this paper applies social network analysis method （SNA） to explore the social network characteristics and conservation strategies of its traditional villages. The results show that: Social networks have the characteristics of high density, good connectivity, stable structure, and small level differences. The number of villages is different in the four blocks, and also the number of receiving and spillover relationships between them varies greatly. The village clusters exhibit a characteristic of "different sizes of subgroups, breaking through county boundaries". The Tongren-Jianzhadong-XunhuaNan cluster located in the southwest has the largest scale, followed by the North Xunhua, Xining, Lanzhou, and Jianzhaxi-Hualong clusters, while the Guide cluster has the smallest scale. The numerical differences in the three types of centrality are significant, with the top 20% of villages ranked in degree centrality and proximity centrality concentrated in the southwest, while the top 20% of villages ranked in intermediary centrality are also distributed in the north and east. The Hehuang region could construct an integrated conservation pattern of "one core, three hearts, and three belts" based on four blocks, important rivers and roads, and conservate six traditional village clusters at three levels, as well as provide different conservation measures for core villages, key villages, and general villages.