Background: Ethical conflict in nursing is a common phenomenon in intensive care units (ICUs). Nurses' ethical sensitivity forms the basis for identifying ethical conflicts. Ethical decision-making abilities are closely related to ethical conflict. However, there are currently no reports on the pathways between ethical sensitivity, decision-making ability, and conflicts among ICU nurses. Purpose: Based on the cognitive-behavioral theory, a structural equation model was developed to quantitatively analyze the relationships between ICU nurses' ethical sensitivity, decision-making ability, and conflicts. Methods: A cross-sectional survey was conducted involving ICU nurses from six general hospitals in China from May to July 2024, using the General Information Questionnaire, Ethical Conflict Nursing Questionnaire-Critical Care Version (ECNQ-CCV), Chinese Moral Sensitivity Questionnaire-Revised Version (MSQ-R-CV, including the dimensions of moral responsibility and strength and sense of moral burden), and Chinese version of judgment about nursing decision (JAND-CE). Descriptive analyses were conducted with SPSS 25.0, and a structural equation model (using Amos 26.0) was performed to identify path relationships between the variables. Results: The constructed model demonstrated a strong overall fit, and there were significant correlations between ethical sensitivity, decision-making ability, and conflicts among Chinese ICU nurses (p < 0.05). The values of path coefficients showed that moral responsibility and strength have a positive association with JAND-CE (beta = 0.263, p < 0.05) and negative association with ECNQ-CCV (beta = -0.246, p < 0.05). Moreover, sense of moral burden has a negative association with JAND-CE (beta = -0.353, p < 0.05) and positive association with ECNQ-CCV (beta = 0.232, p < 0.05). Further, JAND-CE has a negative association with ECNQ-CCV (beta = -0.183, p < 0.05). This study conducted mediation analysis by examining the indirect path between moral responsibility and strength, sense of moral burden, and ECNQ-CCV via JAND-CE, whereby the beta coefficients of independent mediating and mediating-dependent variables were multiplied. The indirect path between moral responsibility and strength and ECNQ-CCV through JAND-CE was significant (i.e. indirect path (0.263 x (-0.183)) = -0.048, p < 0.05, LL = -0.608, UL = -0.07), and the indirect path between sense of moral burden and ECNQ-CCV through JAND-CE was significant (indirect path ((-0.353) x (-0.183)) = 0.065, p < 0.05, LL = 0.082, UL = 0.758) and did not contain a zero value between lower and upper boundaries. Conclusions: This study reveals the dual-path mechanism of moral responsibility and strength and sense of moral burden on ethical conflicts through structural equation modeling, emphasizing the mediating pivotal role of ethical decision-making ability. The research findings provide a theoretical basis for the refinement of moral capacity cultivation systems, while also warning of the potential negative impacts of moral burden. Implications for nursing managers: Nursing managers should dynamically evaluate ICU nurses' ethical sensitivity and decision-making abilities to provide a reference for implementing individualized ethical conflict intervention measures.