Background: Resources are spent worldwide on the development of digital platforms and their user interfaces (UIs) for digital mental health services (DMHS). However, studies investigating the potential benefits of different UIs for DMHS are currently lacking. To fill this knowledge gap, the aim of this study was to investigate differences in treatment engagement between two different UIs for DMHS. Methods: A total of 397 participants from the Swedish general public were randomized (1:1) to use a self-guided digital problem-solving intervention during one week, either with an optimized UI (N = 198), based on user experience (UX) design principles and with automated features, or a basic UI (N = 199), analogous with a UI used in Swedish regular health care comprising elementary UI features and less automation. Primary outcome measures were self-rated usability, on the System Usability Scale, and treatment credibility, on the Credibility/Expectancy Questionnaire. Secondary outcome measures included behavioral engagement with the intervention. Findings: There were no significant differences between the UIs in ratings of usability or treatment credibility. However, participants who used the optimized UI were significantly more engaged with the intervention as measured by usage of the intervention at least once (odds ratio 2.54, 95% CI [1.67, 3.85]), total number of generated solutions (mean difference 1.41, 95% CI [0.72, 2.11]), and mean number of generated solutions per initiated problem-solving attempt (mean difference 1.45, 95% CI [1.06, 1.85]). Other findings included participants using the optimized UI rating the intervention as easier to understand, while feeling more overwhelmed, than those using the basic UI. Interpretation: Our findings indicate that an optimized UI based on UX design principles, in comparison to a basic UI comprising elementary UI features, do not affect overall self-rated usability or treatment credibility but increases some measures of behavioral engagement with a digital intervention.