Background <p>There is a wealth of nursing literature about nursing students’ loneliness, self-efficacy, and influence on academic performance; however, the mediating role of social media addiction among these associations remains scant after the pandemic and in a Middle Eastern context.</p> Objectives <p>To investigate the mediating effect of social media addiction in the interrelationships between loneliness, academic self-efficacy, and academic performance among nursing students.</p> Methods <p>Cross-sectional and correlational study design was used. Five nursing colleges in Saudi Arabia were selected as the study settings using convenience sampling. Nursing students (<i>n</i> = 775) were consecutively invited to participate and completed three standardized self-report scales (6-item Loneliness Scale, 6-item Bergen Social Media Addiction Scale, and the 5-item General Academic Self-Efficacy Scale). For data analysis, the covariance-based structural equation modeling was utilized. Data were analyzed using covariance-based structural equation modeling with maximum likelihood estimation. Path analysis was conducted to test direct and mediating relationships, and model fit was evaluated using <i>χ</i>^<i>2</i>/df, root mean square error of approximation, comparative fit index, Tucker-Lewis index, and standardized root mean square residual.</p> Results <p>Loneliness was negatively associated with social media addiction (<i>β</i> = − 0.18, <i>p</i> = 0.001) and positively associated with academic self-efficacy (<i>β</i> = 0.14, <i>p</i> = 0.001) and academic performance (<i>β</i> = 0.14, <i>p</i> &lt; 0.05), contrary to the hypothesized directions. Social media addiction was positively associated with academic self-efficacy (<i>β</i> = 0.19, <i>p</i> = 0.001) but negatively associated with academic performance (<i>β</i> = − 0.11, <i>p</i> &lt; 0.05). Social media addiction significantly mediated the indirect effect of loneliness on academic self-efficacy (<i>β</i> = − 0.03, <i>p</i> = 0.005), but not on academic performance. Loneliness explained 3.33% of the variance in social media addiction, while loneliness and social media addiction explained 4.57% and 3.89% of the variance in academic self-efficacy and academic performance, respectively.</p> Conclusion <p>Some findings diverged from theoretical expectations, suggesting that the relationships among loneliness, social media addiction, academic self-efficacy, and academic performance may be context-dependent. Social media addiction mediated only the relationship between loneliness and academic self-efficacy. Nursing educators and policymakers should interpret these findings cautiously and consider multifactorial strategies that address loneliness, problematic social media use, and broader academic and psychosocial support needs among nursing students.</p> Clinical trial number <p>Not applicable.</p>