<p>Conventional longitudinal and cross-sectional network analyses have produced inconsistent conclusions regarding the relationship between loneliness and anxiety. However, cross-lagged panel network analysis offers a clearer framework for understanding how these symptoms develop over time. The present study utilized a cross-lagged panel network approach to analyze the longitudinal network associations between loneliness and anxiety in non-clinical adolescents. A total of 816 Chinese adolescents (77.30% female; <i>M</i>age = 16.89, <i>SD</i> = 0.72) were followed over six months. Loneliness and anxiety were measured using the short-form UCLA and the GAD-7, respectively. Results indicated a notable decline in reported loneliness between T1 and T2, with no significant change in anxiety. Loneliness predicted anxiety, whereas anxiety did not predict loneliness. “Yearning for Company” was the most significant predictor of other symptoms, while “Socially Engaged” minimally influenced loneliness and anxiety. These findings suggest that interventions aimed at reducing loneliness, particularly by addressing adolescents’ “yearning for company”, could play a key role in mitigating anxiety and enhancing overall mental well-being in this population. Unlike prior work that has predominantly focused on social anxiety, this study centers on generalized anxiety, a broader and comparatively underexplored form of anxiety that is highly relevant to adolescent development, thereby extending understanding of loneliness–anxiety comorbidity in Chinese adolescents.</p>