The potassium (K) content of New Zealand ryegrass/white clover pastures can exceed the requirements for optimum pasture production (2.5–3.0%) and dairy cow requirements (0.2–0.6%) during late spring. Experiments were carried out at 13 sites in three different dairying regions over 2–3 years on different soil types to determine the effects on the cow's preference for pasture of either increasing pasture K content from applications of 0, 200, 400, and 800 kg of potassium chloride (KCl) ha–1, or the effect of dusting sodium chloride (NaCl) at 0, 5, 10, and 20 kg ha–1 immediately before grazing. Cow preference was assessed by measurements of bite rate, grazing time, pasture dry matter (DM) disappearance, and bite size. In addition, the pasture intake of two groups of cows, one group grazing pasture with K content in the optimum range for pasture production, and one group on pasture above that range, was measured using the alkane technique. In South Taranaki on Allophanic soils, increasing pasture K content significantly decreased mean cow bite rate and grazing time in October 2000 (P=0.006 and 0.001, respectively) and 2001 (P = 0.006 and 0.045, respectively), but not in November 2000 and 2001. Increasing pasture K content also significantly decreased pasture disappearance in October and November 2001, but had no significant effect on bite size. Pasture K content had no significant effect on bite rate or grazing time on brown and podzol soils in Northland in October 2001, but there was a significant positive linear trend (P = 0.011 and 0.039, respectively) in November 2001. There was no significant effect of pasture K content on pasture disappearance and bite size in Northland. Measurement of pasture preference on pumice soils on the Central Plateau in 2002 also showed no significant effect of pasture K content. In South Taranaki, pasture intake was significantly higher (P < 0.05) (18.0 versus 15.3 kg DM cow–1 day—1) for cows grazing pasture with 4.0 versus 3.3% K in October 2002, but there was no significant difference in intake between pastures containing 3.1 and 4.7% K in November 2002. Sodium chloride applied to pasture prior to grazing significantly increased cow bite rate in October and November 2002 (P=0.045 and 0.018, respectively) and grazing time in November 2002 (P = 0.032) on pumice soils. Rate of KCl or NaCl had no effect on pasture nutritive quality with overall organic matter digestibility ranging from 77 to 85%, crude protein from 9 to 26%, neutral detergent fibre from 30 to 51%, and metabolisable energy from 9 to 12 MJ kg DM–1. These results indicate that pasture K had no consistent measurable effect on cow preference, but at some sites in some years there was a trend for cow bite rate and grazing time to be depressed as pasture K content increased above 3.0%. Over a wide range of pasture Na contents (0.03–0.29%), no effect on cow preference was detected from dusting pasture with NaCl just prior to grazing.