Objective: Non-suicidal self-injury (NSSI) is a major public health issue. Despite increased research efforts in clinical samples, we still have little understanding of the short-term correlates, predictors, and targets of NSSI among treatment-seeking individuals. The current study was designed to (a) evaluate how suicidal thoughts, interpersonal difficulties, hopelessness, and affective states are associated with same-day and next-day NSSI; (b) identify which factors may be effective targets in treatment through network modelling.Methods: Data from 1,265 psychiatric inpatients who self-injured throughout their visit to a psychiatric hospital self-reported their suicidal ideation, negative affect, and positive affect on a daily basis (in total 36,345 prospective reports). An additional 632 patients were also surveyed regarding feelings of hopelessness, wish to live, and interpersonal difficulties. Using multilevel structural equation modelling, we examined contemporaneous and time-lagged associations with NSSI. Multilevel network analyses assessed interconnectedness of daily predictors, and were compared with a matched sample of 1,265 patients who did not self-injure during their stay. Results: Increases in suicidal ideation were associated with increased probability of same-day and next-day self-injury, and an inverse relationship was observed for wish to live. Increases in positive affect were also significantly associated with decreased probability of next-day self-injury. Perceived burdensomeness had high centrality in network models, particularly among patients who self-injured, indicating it is susceptible to activation and directly associated with all predictors.Conclusions: Routine monitoring may improve prediction of when a patient is at short-term risk to self-injure and provides person-specific data that can assist in targeting risk and protective factors during treatment.