Microinjection of spermatozoa or spermatids into oocytes is a major choice for infertility treatment. However, the use of premeiotic spermatocytes has never been considered because of its technical problems. Here, we show that the efficiency of spermatocyte injection in mice can be improved greatly by reducing the size of the recipient oocytes. Live imaging showed that the underlying mechanism involves reduced premature separation of the spermatocyte's meiotic chromosomes, which produced much greater (19% vs. 1%) birth rates in smaller oocytes. Application of this technique to spermatocyte arrest caused by STX2 deficiency, an azoospermia factor also found in humans, resulted in the production of live offspring. Thus, the microinjection of primary spermatocytes into oocytes may be a potential treatment for overcoming a form of nonobstructive azoospermia caused by meiotic failure.