Background: A written action plan (WAP) for managing asthma exacerbations is recommended. Objective: We aimed to compare the effect on unscheduled medical contacts (UMCs) of a digital action plan (DAP) accessed via a smartphone web app combined with a WAP on paper versus that of the same WAP alone. Methods: This randomized, unblinded, multicenter (offline recruitment in private offices and public hospitals), and parallel-group trial included children (aged 6-12 years) or adults (aged 18-60 years) with asthma who had experienced at least 1 severe exacerbation in the previous year. They were randomized to a WAP or DAP+WAP group in a 1:1 ratio. The DAP (fully automated) provided treatment advice according to the severity and previous pharmacotherapy of the exacerbation. The DAP was an algorithm that recorded 3 to 9 clinical descriptors. In the app, the participant first assessed the severity of their current symptoms on a 10-point scale and then entered the symptom descriptors. Before the trial, the wordings and ordering of these descriptors were validated by 50 parents of children with asthma and 50 adults with asthma; the app was not modified during the trial. Participants were interviewed at 3, 6, 9, and 12 months to record exacerbations, UMCs, and WAP and DAP use, including the subjective evaluation (availability and usefulness) of the action plans, by a research nurse. Results: Overall, 280 participants were randomized, of whom 33 (11.8%) were excluded because of the absence of follow-up data after randomization, leaving 247 (88.2%) participants (children: n=93, 37.7%; adults: n=154, 62.3%). The WAP group had 49.8% (123/247) of participants (children: n=45, 36.6%; mean age 8.3, SD 2.0 years; adults: n=78, 63.4%; mean age 36.3, SD 12.7 years), and the DAP+WAP group had 50.2% (124/247) of participants (children: n=48, 38.7%; mean age 9.0, SD 1.9 years; adults: n=76, 61.3%; mean age 34.5, SD 11.3 years). Overall, the annual severe exacerbation rate was 0.53 and not different between the 2 groups of participants. The mean number of UMCs per year was 0.31 (SD 0.62) in the WAP group and 0.37 (SD 0.82) in the DAP+WAP group (mean difference 0.06, 95% CI -0.12 to 0.24; P=.82). Use per patient with at least 1 moderate or severe exacerbation was higher for the WAP (33/65, 51% vs 15/63, 24% for the DAP; P=.002). Thus, participants were more likely to use the WAP than the DAP despite the nonsignificant difference between the action plans in the subjective evaluation. Median symptom severity of the self-evaluated exacerbation was 4 out of 10 and not significantly different from the symptom severity assessed by the app. Conclusions: The DAP was used less often than the WAP and did not decrease the number of UMCs compared with the WAP alone.