Six mid‐lactation multiparous Holstein–Friesian dairy cows were used to examine the potential of a fermented whole‐crop barley (Hordeum vulgare)/kale (Brassica oleracea) bi‐crop as a feed compared with a first‐cut perennial ryegrass silage. The barley/kale bi‐crop was grown as a strip intercrop, and was harvested and ensiled as an intimate mixture [0·80 barley and 0·20 kale on a dry‐matter (DM) basis]. Animals were offered ad libitum access to one of three experimental diets in a duplicated Latin Square design experiment: (i) Bi‐crop (the barley/kale bi‐crop); (ii) Grass (the grass silage); and (iii) Mix (a 1:1 fresh mixture of Bi‐crop and Grass). All animals also received a standard dairy concentrate at a rate of 4 kg d−1 in equal portions at each of two milkings. The Bi‐crop and Grass silages contained 346 and 293 g DM kg−1, 108 and 168 g crude protein kg−1 DM, 268 and 36 g starch kg−1 DM, and had pH values of 3·87 and 3·80 respectively. Animals offered the two bi‐crop silage‐containing diets consumed more forage DM than those offered grass silage (14·6, 14·9 and 12·6 kg DM d−1 for Bi‐crop, Mix and Grass respectively; s.e.d. 0·45, P < 0·01) and yielded more milk (24·0, 23·9, 22·6 kg d−1 for Bi‐crop, Mix and Grass respectively; s.e.d. 0·26, P < 0·01). However, differences in the partitioning of dietary nitrogen towards milk protein and away from excretion in urine suggest a more efficient (rumen) utilization of feed protein by animals offered diets containing the bi‐crop silage. It is concluded that, despite having a low crude protein concentration, barley/kale bi‐crop silage offers excellent potential as a feed for lactating dairy cows.