Non pregnant patients with pyelonephritis have been treated successfully as outpatients. The purpose of this study was to compare the safety and efficacy of outpatient and inpatient treatment of acute pyelonephritis in pregnancy.This was a randomized controlled, clinical trial. One hundred twenty eight gravidas past 24 weeks' gestation admitted in two large teaching hospitals in Iran, divided by random blocks to outpatient or inpatient therapy, received intramuscular ceftriaxone at 24-hour intervals while hospitalized, then were discharged and reevaluated within 48–72 hours or remained hospitalized until afebrile for 48 hours. All patients completed a 14-day course of oral cephalexin. There were not any significant differences between inpatients and outpatients group about the response to initial treatment (14.1% vs. 6.3 %), recurrent pyelonephritis (10.9% vs. 25%), and gestational age at delivery (37.2 ± 1.6. vs. 38.9 ± 0.2). The Persistent bacteriuria (46.9% vs. 17.2%) and preterm deliveries (18.8% vs. 3.1%) in inpatients were significantly more than outpatients. Recurrent bacteriuria was significantly more in outpatients (53.1% vs. 21.8%). The birth weight of neonate in inpatients (3120.2 ± 463) was significantly more than outpatients (2659 ± 382.3). Although outpatient antibiotic therapy is effective and reasonable in selected pregnant women with pyelonephritis. Because of a high rate of low birth weight and recurrent bacteriuria in outpatient group in our study, a selective approach to outpatient management of acute pyelonephritis in pregnancy at later gestations should be accomplished.