Teletherapy offers new opportunities to address attendance challenges in adults with attention-deficit/hyperactivity disorder (ADHD). This quasi-experiment involved primary, prospective analyses of 15-week teletherapy attendance data from adults with ADHD, focusing on digital engagement and predictive modeling. Participants showed a female-to-male ratio of 3:1, with 12.8% identifying as trans men and 25.5% as non-binary, reflecting inclusivity of gender nonconforming, treatment-seeking clients. Average teletherapy attendance was 85.4% (standard deviation [SD] = 10.3%), significantly higher than prior in-person attendance of 72.3% (SD = 12.4%; t[46] = 5.21; p<0.001; d = 1.15). Attendance differed by ADHD subtype (p=0.027), with inattentive individuals showing the highest engagement. Attendance correlated positively with automated reminders (β = 0.45; p<0.01), self-scheduling (β = 0.38; p<0.05), and reminder acknowledgment (r = 0.41; p=0.006), and negatively with mobile-only device use (r = –0.33; p=0.032; odds ratio [OR] = 0.52, p=0.041). Logistic regression identified automated reminders as a strong predictor of adherence (OR = 3.4, p=0.004). Random forest classification achieved 87% accuracy (F1 = 0.86; area under the precision–recall curve = 0.89), with reminders, login latency, and age as top features. Interaction analyses showed subtype-specific effects: reminders benefited inattentive users (β = 0.52, p=0.009), and self-scheduling improved attendance among participants with combined subtypes (β = 0.33, p=0.041). Age modestly correlated with attendance (r = 0.32, p=0.045) but was nonsignificant in multivariate models. Sensitivity analyses confirmed robustness. Findings emphasize digital engagement and behavioral data as superior predictors of adherence compared to demographics, underscoring teletherapy’s potential for neurodivergent adults across diverse gender identities.