Along with the tremendous reconfiguration of social space in urban China, the place attachment of urbanites to their community or the city may undergo transformation or restructuration, yet has been rarely touched by existing literature. In particular, the measurement and mechanism of the place attachment deserve high attentions, and there is in need of an in-depth empirical study. In this vein, with a case of a typical large city, Guangzhou, this paper explores the dynamics of residents' place attachment, both theoretically and empirically. In details, we first examine the role of various forces such as social dimension (framed by social trust), functional dimension (framed by residential satisfactory to the services and facilities in the neighborhood), and their interactions underlying the making of place attachment. Second, we examine the interrelation between residents' place attachment to various scales, community and city. Third, we compare the place attachment for two groups, local residents and migrants, with respect to the extent of their place attachment and mechanisms. The empirical study includes 23 communities surveyed in Guangzhou. Our sampling strategy, questionnaire and resultant database prove to be representative. It includes the measurement of the degree of place attachment of residents through known Likert scale; then we apply ANOVA analysis by SPSS21.0 and Structural Equation Model by STATA14.0 to test the effects of various factors, such as social trust and residential satisfaction, for different types of urbanites. Our empirical study reveals that: first, there are two types of place attachments: community attachment and city attachment. The level of urban residents' community attachment is higher than that of city attachment. Second, the level of local residents' place attachment is significantly higher than that of migrants. Third, urban residents' community attachment is significantly affected by both social trust and residential satisfaction, but without the mediating effect of community attachment these two dimensions do not have direct impacts on city attachment. Fourth, social trust and residential satisfaction have significant effects on both community attachment and city attachment of migrants; but their impacts on local residents' city attachment are relatively weak. Our findings indicate that the mechanism of place attachment is marked by both spatiality and heterogeneity. For instance, for different spatial scales, the determinants of place attachment and their effects are by no means the same. Moreover, there are varied mechanisms of place attachments for different types of residents or social groups. As such, we call for policy concerns upon the differentiation of stratified social groups and spatial contexts, when China implements its strategy to promote people-centered urbanization and social wellbeing, or against the context of developing countries or transitional economy.