Loneliness is a common public health problem that influences people's physical and mental health. There is a high incidence of loneliness in adolescents. Some research suggested that smartphone or Internet addiction (SA or IA) may be a factor. But the relationship between loneliness and SA or IA is not completely clear among adolescents. We aim to estimate the correlation coefficient r between them. Databases, consisting of PubMed and Web of Science, were retrieved systematically for studies of the association between adolescents' loneliness and SA or IA. The Newcastle-Ottawa Scale was chosen as an assessment tool in this analysis. We estimated the correlation coefficient r between loneliness and SA or IA and drew a forest plot. Moreover, moderator analyses were also conducted to explore what leads to heterogeneity in our study. 21 studies were finally included in our analysis with 27,843 samples. The pooled correlation coefficient r was 0.252 (95% confidence interval: [0.173, 0.329]; < 0.001) with low heterogeneity (I = 0.000%; Q = 23.616; < 0.001), indicating a moderate positive association. The funnel plot indicated small publication bias. A one-study removal sensitivity analysis indicated there was no significant difference between these studies. Meta-regression indicated no significant difference between the results and age (Q = 11.94, df = 18, = 0.8504). Our analysis indicated a moderate positive association between loneliness and SA or IA. The results may attract the attention of some experts who study adolescent psychological problems and behavioral problems and may provide ideas for their research in the future.