AimsTo explore reciprocal relationships between real-time gambling cravings and self-efficacy with gambling behaviour, and the moderating role of gambling, mental health, and addiction-related variables. DesignSecondary analysis of a 4-week Ecological Momentary Assessment (EMA) study conducted in Tasmania, Australia. SettingData were collected via telephone interviews (pre-EMA) and smartphones (EMA). ParticipantsNinety-seven regular gamblers (mean age: 45.90 years, 57.73% male) reported 5,113 observations. MeasurementsEMA measures included gambling cravings (occurrence, frequency, intensity), self-efficacy (craving-related, gambling-related), and gambling behaviour (episodes, expenditure, duration). Pre-EMA measures included gambling (severity; harms; motives; high-risk situations), mental health (depressive symptoms; anxiety symptoms) and addiction-related (alcohol use; smoking; substance use) moderator variables. FindingsMixed-effects binary logistic regression analyses revealed that gambling cravings predicted gambling episodes (OR = 2.23, 95% CI:1.61, 3.08), gambling self-efficacy and gambling duration were reciprocally related (OR = 4.65, 95% CI:1.08, 20.04; OR = 0.21, 95% CI:0.05, 0.93), and craving self-efficacy predicted gambling expenditure (OR = 0.30, 95% CI:0.10, 0.86). Moderation analyses revealed that: (1) craving self-efficacy exacerbated craving frequency with gambling expenditure; (2) coping motives exacerbated gambling self-efficacy with gambling expenditure; (3) high-risk positive reinforcement situations exacerbated craving intensity and gambling self-efficacy with gambling episodes, and gambling episode with craving occurrence; and (4) substance use exacerbated gambling self-efficacy with duration, and buffered gambling expenditure with craving intensity. ConclusionsThese findings have implications for the development of real-time gambling interventions that aim to reduce gambling cravings and increase self-efficacy, which could be targeted to vulnerable individuals, including people who frequently gambled for coping purposes or positive reinforcement, and people with comorbid substance use.