The cost of healthcare in the U.S. Has been rising with an alarming rate during the last two decades. One of the causes for the rising cost is medical bad debt. There is surprisingly very little research that explores the performance of computational intelligence and soft computing methods in improving bad debt recovery in the healthcare industry. This study examines the performance of an adaptive neuro-fuzzy inference system (ANFIS) with semi-supervised learning (SSL) in classifying bad debt in the healthcare context, as better debt classification leads to greater recovery. Computer simulation shows that ANFIS with SSL is a viable method. Our model generated better classification accuracy than ANFIS alone did and our accuracy is better than those in the previous studies. Insightful interpretation of the results is provided through data clustering and analysis of control surfaces. The latter depicts nonlinear interaction between various factors contributing to bad debt. Additional analysis is provided through receiver operating characteristic (ROC) charts to interpret the classification accuracy rates at a continuum of cut-off points from within the range [0, 1]. These results and their analysis show the potential of ANFIS with SSL models in classifying unknown cases, which are a potential source of revenue recovery.