By means of the results of In vitro fertilization(IVF) of supernumerary oocytes, the possibility of predicting a pregnancy outcome following gamete intrafallopian transfer(GIFT) and the prognostic implications for future pregnancy were evaluated in 46 GIFT cycles excluding simultaneous program of GIFT and IVF from February, 1989 to July, 1991. IVF of supernumerary oocytes were identified in 21 cycles, but not in remaining 25 cycles. There was no significant difference in age, duration and etiologic factors of infertility, and serum levels of FSH, LH and $E_2$ on MCD #3 and $E_2$ on the day of hCG injection between fertilized(N=21) and unfertilized group(N=25). The number of oocytes retrieved was similar in both groups. The number of supernumerary oocytes available for IVF after transfer was $5.43{pm}2.95$ ranging from 2 to 12. The prenancy rate in fertilized group, 33.3%(7/21), was higher without statistical significance, compared with 8.0%(2/25) in unfertilized group. Using IVF of supernumerary oocytes as a test of pregnancy following GIFT, sensitivity was 77.8 %; specificity, 61.2%; positive predictive value(PPV), 33.3%; negative predictive value(NPV), 92%. The fertilization rate of supernumerary oocytes in pregnant group, $86.4{pm}22.8%$ was significantly higher compared with $56.1{pm}20.2%$ in nonpregnant group. In cases with fertilization rate ${geq}80%$, pregnancy was expected with PPV of 85.7%. In conclusion, IVF of supernumerary oocytes in GIFT program can be a profitable method as a prognostic indicator of pregnancy following GIFT. More aggressive diagnostic and therapeutic measures should be performed in cases with failure in IVF of supernumerary oocytes.