Background Lack of physical activity and excessive sitting time contributed to ectopic fat accumulation, especially in the liver. Previous studies have illustrated the harm of sedentary behaviour and the benefits of physical activity on fatty liver disease. We aimed to explore the association between the behaviour patterns and the risk of metabolic dysfunction-associated fatty liver disease (MAFLD) using isotemporal substitution model to examine the effect of replacing one behaviour to another while keeping the total time and other behaviours fixed among Chinese middle-aged and elderly population. Methods This study included 161 147 participants aged >= 40 years old from the nationwide, population-based cohort of the REACTION study. The International Physical Activity Questionnaire was used to measure self-reported time for sleeping, sitting, walking and moderate-to-vigorous physical activity (MVPA). MAFLD was defined by evidence of fatty liver index (FLI) >= 60 in addition to one of the following three patterns, namely overweight/obesity, presence of diabetes, or evidence of metabolic dysregulation. Isotemporal substitution models using logistic regression models to evaluate the association of replacement of different behaviour patterns with each other and the risk of MAFLD. Results Substitution of 60 minutes per day of sleeping, walking or total MVPA for sitting was associated with a 2%-8% reduction of MAFLD risk in overall participants. In employed individuals, replacing sitting time with occupational MVPA or nonoccupational MVPA both could bring benefits to liver steatosis. Stratified analysis found that replacing 60 minutes of sitting time with an equivalent time of other behaviour pattern could reduce approximately 8% of the risk among MAFLD participants with metabolic abnormalities. Such a relationship might be explained by the important mediated role of metabolic elements, such as waist circumference, body mass index, triglycerides and homoeostasis model assessment of insulin resistance. Furthermore, replacing sitting with MVPA showed a stronger association among participants who got enough sleep (sleep duration >= 7 hours per day). Conclusion Replacing sitting with other behaviour patterns could reduce the prevalence of MAFLD, and such substitution effect was much remarkably in individuals with abnormal metabolic status. Observably, obese individuals were more likely to benefit from appropriate changes in behaviour patterns. Moreover, the analysis of sleep duration stratification appealed that the adequacy of individual sleep duration also had a significant impact on the substitution effect. It is worth noting that adjusting the time allocation of behaviour patterns might have a beneficial impact on liver-metabolic health, and these findings might help us better recognize the importance of reasonable arrangement of behaviour patterns according to the individual's situation.