<p>Early intervention of personality disorders (PD) in adolescents is crucial due to their long-term developmental impact. Dimensional models, like the DSM-5 Alternative Model for Personality Disorders (AMPD), may facilitate early detection and intervention, although their applicability in adolescent clinical populations is under-researched. This longitudinal study examines changes in PD diagnosis and impairments in personality functioning according to the AMPD and explores potential predictors in a clinical adolescent cohort. 227 adolescents (<i>M</i> age = 15.25, 85.02% female) from an outpatient clinic for early PD intervention were assessed at baseline, one year, and two years. Personality functioning was measured using the Semi-Structured Interview for Personality Functioning, psychiatric diagnoses with the Mini International Neuropsychological Interview, and psychosocial functioning with the Social and Occupational Functioning Assessment Scale. Mixed regression models were applied for analysis. The probability of a PD diagnosis significantly decreased over two years (OR = 0.54, p = .011). Significant improvements were observed in identity (β = -0.191, <i>p</i> &lt; .001), self-direction (β = -0.243, <i>p</i> &lt; .001), empathy (β = -0.147, <i>p</i> &lt; .001) and intimacy (β = -0.136, <i>p</i> &lt; .001). Age, sex, psychiatric comorbidity, psychosocial functioning at baseline, treatment setting (inpatient versus outpatient) and dose did not influence PD diagnosis remission or changes in personality functioning, but higher baseline psychosocial functioning was linked to less improvement in self-direction (β = 0.008, <i>p</i> = .033). Personality functioning improved over time, supporting the notion that PD pathology in adolescents is amenable to change. Further research is needed to examine individual trajectories of personality functioning and refine targeted intervention strategies.</p>