Relaxation training with and without muscle contraction was evaluated in 30 adult patients undergoing biofeedback for psychophysiological disorders. Participants were randomized to two sessions of relaxation training with the type of training randomly determined for the first test session. Surface electromyographic activity was significantly lower during relaxation with muscle contraction (M ± SD: 4.68 ± 2.92 µV) compared to relaxation without muscle contraction (M ± SD: 5.91 ± 4.07 µV) (t(29) = 2.44, p = .021). Participants were about twice as likely to report that they preferred relaxation training without muscle contraction (53.3%) over training with muscle contraction (23.3%). No significant differences were found for fingertip temperature. The implications for the use of relaxation training in clinical practice are discussed.