AIMS: It is unclear whether there is a causal relationship between social determinants of health (SDoH) and mortality in patients with chronic heart failure (CHF). In this study, we assessed causality using inverse probability weighting (IPW) of marginal structural models (MSMs) during the course of CHF. METHODS AND RESULTS: A total of 1377 patients with CHF were enrolled in this multi-centre, prospective cohort study from September 2017 onwards. The social domain and two dimensions of the chronic heart failure patient-reported outcomes measure (CHF-PROM) were used to assess SDoH, social support, and support utilization in these patients. CHF-PROM and mortality information were obtained at 1, 3, and 6 months following discharge and every 6 months thereafter at regular follow-up visits. The impact of SDoH, social support, and support utilization on mortality was analysed by logistic regression and IPW of MSMs. Logistic regression showed that SDoH, social support, and support utilization at baseline were not associated with mortality. After adjustment for confounders, MSMs showed that SDoH and social support were not associated with mortality at baseline. In contrast, low support utilization at baseline and unfavourable SDoH, low social support, and low support utilization during follow-up increased the risk of death. CONCLUSION: Using follow-up data and analysis of MSMs, we found that long-term out-of-hospital effects of SDoH, but not one-time effects, were risk factors for mortality in patients with CHF. SDoH should be assessed during the entire course of CHF to prolong patient survival. REGISTRATION: Chinese Clinical Trial Registry, ChiCTR2100043337 (https://www.chictr.org.cn/showproj.html?proj=64980).