Abstract Background Patients, who discontinue early, do not benefit from phase I/II clinical trials (early-phase clinical trials (EPCT)). In this study, associations between objective smartphone measurements of physical activity and fitness and early trial discontinuation in patients with cancer participating in EPCT were investigated. Methods Before start of treatment, physical activity (steps/day) and physical fitness (meters walked in 6 min) were measured with a smartphone, and patient-reported physical function (PRO-PF) was assessed (EORTC QLQ-C30-PF). Early trial discontinuation was defined as discontinuation ≤ 28 days. Univariable logistic regression analyses were performed to study associations of physical activity, fitness, and function with early trial discontinuation. Optimal cutoff values of physical activity and fitness were assessed with ROCs, based on positive predictive values (PPV). Results Median (interquartile range (IQR)) step count was 4263 (2548–6897) steps/day, mean ± standard deviation 6-min walking distance was 477 ± 120 m and median (IQR) PRO-PF score was 83 (67–95) points. Fourteen patients (12%) discontinued the trial early. Smartphone measurements of physical activity in units of 100 steps per day (odds ratio (OR) = 0.96, 95% CI = 0.94–0.99, p = 0.01), physical fitness (OR = 0.99, 95% CI = 0.98–0.99, p < 0.01), and PRO-PF (OR = 0.97, 95% CI = 0.94–1.00, p = 0.03) were associated with early trial discontinuation. Optimal cutoff values were < 900 steps for physical activity and < 285 m for physical fitness. PPV for early trial discontinuation was 100% in patients who walked both < 1500 steps per day and < 300 m in 6 min. Conclusions Objective smartphone measurements of physical activity and fitness are associated with early trial discontinuation. However, cutoff values should be externally validated in a larger cohort before implementation in clinical practice.