This study compared the reproductive performance of sows that were submitted to cervical artificial insemination (CAI) using concentrations of 3.5 x 109 spermatozoa per dose or intrauterine artificial insemination (IUAI) with 2, 1, or 0.5 x 109 spermatozoa per dose. Within all treatments, females were inseminated either after conventional estrus detection or after ultrasound-guided ovulation diagnosis. Farrowing rate did not differ between CAI and IUAI, regardless of the concentration of spermatozoa used (P > 0.05). In comparison with CAI in parity-3+ females, the probability of failure to farrow was greater (P < 0.05) with IUAI in parity-1 females inseminated with 2 x 109 spermatozoa per dose, parity-2 females with any concentration, and parity-3+ females inseminated with 0.5 x 109 spermatozoa per dose (P < 0.01). Total litter size did not differ across AI methods (P > 0.05), but it was smaller (P < 0.0005) in parity-1 females than for higher parity females. Total litter size was lower with IUAI in parity-1 females with any concentration and in parity-2 females with 2.0 x 109 spermatozoa per dose (P < 0.05). The estimation of ovulation via conventional estrus detection or ultrasound, as well as the occurrence of semen backflow and bleeding during AI, did not influence farrowing rate or litter size (P > 0.05). Intrauterine AI had a lower probability of farrowing with the concentration of 0.5 x 109 spermatozoa per dose and in parity-2 females and also reduced litter size in parity-1 females.