Round spermatid injection (ROSI) into oocytes offers men with non-obstructive azoospermia (NOA) the opportunity to have biological offspring in cases where mature spermatozoa are not detected. However, the clinical outcomes of ROSI remains poor. This study compares the outcomes of ROSI with intracytoplasmic sperm injection (ICSI) and investigates the effect of hormonal pretreatment. This retrospective cohort study enrolled infertile couples undergoing either ROSI or ICSI at the reproductive center in Taipei Veterans General Hospital. The administration of selective estrogen receptor modulators, gonadotropins, and aromatase inhibitors in male patients were collected. Relevant hormonal markers and biochemical parameters were determined. The outcomes of ROSI and ICSI were assessed based on fertilization rate, implantation rate, and live birth rate. A total of 36 couples were recruited in the ROSI group, whereas 39 couples were recruited in the ICSI group for the analysis. Patients in each group demonstrated similar characteristics, except for a higher proportion of male patients in the ROSI group who were pretreated with anastrozole. The fertilization rate and implantation rate were similar between ROSI and ICSI groups after adjusting for confounding variable. The live birth rate was significant lower in the ROSI group (8.3%) than in the ICSI group (30.8%) before and after adjusting for confounding variable. ROSI demonstrated fertilization and implantation rates comparable to those of ICSI for male patient with NOA undergoing testicular sperm extraction surgery. Further studies evaluating the effect of anastrozole administration on ROSI outcomes are warrant.