Abstract: This study compared the effects of eight hours of training in NLP with eight hours of training on the Carkhuff model on the therapeutic relationships of counselors-in-training enrolled in their initial counseling practicum. Research was conducted at a medium size university in the Rocky Mountain region. Research subjects consisted of 46 counselors-in- training. Ratings were made on the Barrett-Lennard Relationship Inventory on each counselor's work with each client. Raters were the counselors, clients, professors who supervised each session and doctoral students who assisted with supervision. Professors and doctoral students also rated each counselor on the Counselor Evaluation Rating Scale. It was hypothesized counselors trained on the Carkhuff model would be rated higher on all the dependent measures by all the raters. Data were analyzed in four separate MANOVAs. None of the null hypotheses were rejected. That being the case, univariate Fs were not computed. Since NLP trained counselors received scores that did not differ at significant levels from those received by Carkhuff model trained counselors, it was concluded that the NLP approach with its emphasis on nonverbal and process techniques to establish a therapeutic relationship worked as well as the established Carkhuff model. Two conclusions were drawn. First, since NLP techniques produced results that approximated those of the Carkhuff model, some NLP techniques can usefully be integrated into the training of counselors. Second, since NLP trained therapists were not rated higher at significant levels than Carkhuff trained counselors, some of the claims of NLP proponents need to be further evaluated. Further research is necessary to replicate this study with a no-treatment control group, to use pre-and post-measures to determine the possible gains in counselor effectiveness as a result of NLP training, and to examine the effects of ongoing supervision in each of the training modalities on the skill levels of counselors-in-training. It was also suggested that in view of the differences in ratings given by supervisor interns and professors on the dependent measures that research be done to examine the intercorrelations of the ratings and determine if different criteria were being used by professors and supervisor interns.