ABSTRACT Background The relationship between the systemic immune–inflammatory index (SII) and the prognosis of cardiovascular disease (CVD) patients with diabetes or prediabetes remains uncertain. This study investigated the association between baseline SII and all‐cause and cardiovascular mortality in American adults with CVD and diabetes or prediabetes. Methods Our survey included 4,060 adults with cardiovascular disease and diabetes or prediabetes from the National Health and Nutrition Examination Survey (1998–2020). Using restricted cubic splines (RCS) based on Cox regression models and a two‐piecewise Cox proportional hazards model for both sides of the inflection point, we elucidated the nonlinear relationship between baseline SII and mortality. Mediation analysis was used to explore the indirect impact of SII on mortality through eGFR. Results In the median 129 months of follow‐up, 620 people died from cardiovascular causes and 1,800 from all causes. In the CVD population with diabetes or prediabetes, SII showed a U‐shaped relationship with all‐cause mortality. The association between SII and CVD mortality was nonlinear and J‐shaped. Stratified and interaction analysis confirmed the stability of the core results. Notably, eGFR partially mediated the association between SII and both all‐cause and cardiovascular mortality by 9.4% and 6.9%, respectively. Conclusions SII revealed a U‐shaped relationship with all‐cause mortality (inflection point: lnSII = 6) and a J‐shaped association with CVD mortality (inflection point: lnSII = 5.73) in CVD patients with diabetes or prediabetes among American patients. Thus, assessing the SII index may offer valuable insights into risk assessment and prognosis in patients with CVD who are diabetic or prediabetic.