Aim: To determine whether minimal stimulation with clomiphene and gonadotropin provides outcomes and direct costs comparable with those of a conventional GnRHa-gonadotropin stimulation protocol for infertile patients undergoing in vitro fertilization. Methods: A non-randomized clinical trial was conducted from 1 July 1996 to 31 March 2003 at the Infertility and Assisted Reproductive Unit, Ramathibodi Hospital, Faculty of Medicine, Mahidol University, Thailand. A total of 192 patients were recruited of whom 96 cases underwent ovarian stimulated cycles with minimal stimulation protocol, and 96 controls underwent ovarian stimulated cycles with GnRHa-gonadotropin protocol, with cases and controls matched for age and infertility cause. Results: The median patient age was 35 years. Endometriosis was the most frequent infertility cause (28.1%). The conventional GnRHa-gonadotropin protocol could give more oocyte numbers than the minimal stimulation protocol (7.3 ±4.9 vs 4.5 ±3.3 oocytes). The fertilization rate and cleavage rate were similar (73.4 ±31.9 and 84.9 ±32.6 in minimal stimulation protocol, 69.3 ±29.6 and 88.4 ±28.0 in GnRHa-gonadotropin protocol, respectively). The pregnancy rate per oocyte retrieval cycle in the GnRHa-gonadotropin protocol was similar to the minimal stimulation protocol. (13.1% vs 13.0%, P =1.000). However, the cost per pregnancy of minimal stimulation protocol was less than that of GnRHa-gonadotropin protocol. ($US 6021.95 for minimal stimulationprotocol per pregnancy, $US 10 785.65 for GnRHa-gonadotropin protocol per pregnancy, P <0.000). Conclusion: Minimal stimulation was less effective than conventional GnRHa-gonadotropin on the ovarian stimulation. However, the total costs of minimal stimulation were cheaper than the conventional GnRHagonadotropin protocol. The decreased costs of minimal stimulation justifies further evaluation of its role in the treatment of infertility in selected cases.