The objective of this experiment was to use diurnal and temporal changes in herbage composition to create two pasture diets with contrasting ratios of water-soluble carbohydrate (WSC) and crude protein (CP) and compare milk production and nitrogen-use efficiency (NUE) of dairy cows. A grazing experiment using thirty-six mid-lactation Friesian x Jersey cows was conducted in late spring in Canterbury, New Zealand. Cows were offered mixed perennial ryegrass and white clover pastures either in the morning after a short 19-day regrowth interval (SR AM) or in the afternoon after a long 35-day regrowth interval (LR PM). Pasture treatments resulted in lower pasture mass and greater herbage CP concentration (187 vs. 171 g kg−1 DM) in the SR AM compared with the LR PM but did not affect WSC (169 g kg−1 DM) or the ratio of WSC/CP (1·0 g g−1). Cows had similar apparent DM (17·5 kg DM cow−1 d−1) and N (501 g N cow−1 d−1) intake for both treatments. Compared with SR AM cows, LR PM cows had lower milk (18·5 vs. 21·2 kg cow−1 d−1), milk protein (0·69 vs. 0·81 kg cow−1 d−1) and milk solids (1·72 and 1·89 kg cow−1 d−1) yield. Urinary N concentration was increased in SR AM, but estimated N excretion and NUE for milk were similar for both treatments. Further studies are required to determine the effect of feeding times on diurnal variation in urine volume and N concentration under grazing to predict urination events with highest leaching risk.