OBJECTIVES: To examine the associations between typology of family relationship and risk of elder mistreatment (EM). DESIGN: Population-based epidemiological cross-sectional study. SETTING: Community-dwelling older adults in the greater Chicago area. PARTICIPANTS: A total of 3157 US older Chinese immigrants. MEASUREMENTS: EM was measured by a brief screening tool. Latent class analysis was used to construct typologies of family relationship from eight indicators, evaluating the structural, associational, functional, affectual, and normative aspects of family relationship. Logistic regression models were used to test the associations between typology of family relationship and EM. RESULTS: Unobligated ambivalent type (44.77%) was the most common family type among US older Chinese immigrants. The prevalence of EM was around 15%, with 475 participants reporting experience of EM. Unobligated ambivalent (odds ratio [OR] = 1.90; 95% confidence interval [CI] = 1.54-2.34) and detached (OR = 1.78; 95% CI = 1.32-2.42) family types were associated with greater risk of EM; tight-knit (OR = .34; 95% CI = .27-.44) family type was associated with lower risk of EM. The relationship between commanding conflicted family type and EM was not significant. CONCLUSION: Unobligated ambivalent family type, featured by high emotional closeness and high family conflict, was prevalent among US Chinese families and associated with EM consequences. Culturally customized social services are suggested to reduce intergenerational ambivalence and promote family harmony for immigrant families. J Am Geriatr Soc 67:S493-S498, 2019.