Background. A randomized controlled trial was conducted to determine the effects of web-based supervisor training on worksite mental health on supervisor support and psychological distress among subordinate workers. Methods. Section chiefs in a computer engineering company were randomly assigned to either a training group (n = 9) or a non-training group (n = 7). The section chiefs in the training group participated in web-based self-learning training on worksite mental health. A total of 92 subordinate workers under the trained section chiefs (the intervention group) and 84 subordinate workers under the untrained section chiefs (the control group) completed a questionnaire at baseline and a 3-month follow-up. Results. The score of supervisor support greatly decreased in the control group during the follow-up period, and the score changed very little in the intervention group, with a significant intervention effect (P = 0.032). This pattern was more pronounced for one particular item dealing with the extent to which a supervisor listens to personal problems of subordinate workers (the intervention effect, P = 0.012). No intervention effect was observed for the score measuring co-worker support, psychological distress, or other job stressors among subordinate workers (P > 0.05). Conclusions. It is suggested that the web-based training of supervisors on worksite mental health is useful in improving, or at least maintaining, supervisor support among subordinate workers. © 2005 Elsevier Inc. All rights reserved.