The purpose of the experiments was to compare two procedures to improve supervisor skills of setting goals, providing feedback, providing systematic prompts, adhering to supervision policies, and using appropriate interpersonal skills with trainees. The first procedure combined a purported contingency of reinforcement with a lecture that reviewed the steps of each skill with written and verbal feedback provided after supervision meetings. The second procedure added workshops based on behavioral skills training (BST). A multiple baseline design across goal setting, feedback, and prompting (supervisory skills) with six supervisors was used. In Experiment 1, the baseline condition improved adherence to supervision policies, but did not teach the supervisory skills to criterion levels. All supervisors demonstrated criterion levels of interpersonal skills during baseline. The workshops taught the supervisory skills, but required individual booster sessions for some skills. During maintenance probes, supervisors demonstrated variable levels of the supervisory skills while maintaining adherence to supervision policies and interpersonal skills. Experiment 2 updated the training based on limitations in Experiment 1. The baseline condition again resulted in adherence to supervision policies, but did not improve the supervisory skills. The workshops were effective in teaching the target skills without booster sessions. During maintenance probes, one supervisor completed all skills at 100% accuracy, whereas the second supervisor left her position before maintenance was evaluated. Both experiments demonstrated a moderate positive correlation between improvement in supervisor performance and trainee performance. The experiments extended the literature on supervision, methods to train trainees, and modifications for BST procedures.