This study sought to compare the efficiency of different nitrogen (N) fertilizer forms applied to perennial ryegrass swards that were rotationally grazed by dairy cows or sheep under Irish conditions for two or three years. A 3 × 2 factorial random complete block design plot arrangement was used to compare calcium ammonium nitrate (CAN), urea and urea + N-(n-butyl) thiophosphoric triamide (urea + NBPT) at 150 and 250 kg N/ha per year. Zero nitrogen plots were also added to allow for N efficiency to be calculated. The study was conducted at four sites, giving three years of data collection at two sites and two years at the other two sites. All four sites observed similar responses to N fertilizer type and N fertilizer rate. Significant differences were observed between the 150 kg N/ha and 250 kg N/ha treatments for pre-grazing herbage yield (1346 and 1588 kg DM/ha, respectively; P < 0.001) and total herbage production (12 290 and 14 448 kg DM/ha, respectively; P < 0.001). There was no difference but a tendency for pre-grazing herbage yield to be higher for CAN and urea + NBPT than urea (1485, 1480 and 1436 kg DM/ha, for CAN, urea + NBPT and urea, respectively; P = 0.091). Total herbage production was significantly higher for CAN and urea + NBPT than urea (13 478, 13 542 and 13 087 kg DM/ha, respectively; P = 0.004). In conclusion, there was an overall benefit detected over the 10 site-years from using urea protected with NBPT v. using urea.