Objective: This two-arm randomized controlled trial aimed to examine the effectiveness of a smartphone-based online platform which incorporates ecological momentary assessments and ecological momentary interventions with the use of machine learning to reduce psychological distress in working adults. Method: 205 working adults were recruited from a local company. Eligible participants were randomized into either the intervention (n = 106) or the waitlist control condition (n = 99). Participants received 4-week smartphone-based personalised psycho-education sessions and evidence-based exercises related to cognitive behavioural and mindfulness principles, driven by data about emotional state, mood, and feedback after each exercise. Reduction in symptoms of depression, anxiety, and stress, was examined as the primary outcome measure. Results: Significant time (before vs. after intervention) by group (intervention vs. control) interaction effect was found, F (1, 175) = 56.67, p < .001, η2= .25. Participants in the intervention group reported less depression, anxiety, and stress symptoms after the intervention (M = 14.38, SD = 15.14) than before the intervention (M = 28.90, SD = 17.99). The attrition rate was 86%. Conclusions: This smartphone-based intervention leveraging both ecological momentary assessments and ecological momentary interventions and employing machine learning was effective in reducing distress in working adults. Findings have informed the next phase of this digital health platform, which will optimize the intervention model via personalization, and contribute to the expanding field of digital mental health.