Four experiments were conducted to evaluate the palatability of sorghum, maize, rye and lupine for pigs. Diets containing sorghum, maize, rye or lupine (test diets) were offered in a series of double choice tests against a reference diet. To prepare the reference diets, white broken rice in the experiments with cereals, and SBM-56 (a soy protein product low in anti-nutritional factors) in the experiment with lupine, were used as the cereal and the protein source of reference, respectively. Six test diets containing sorghum, maize or rye were prepared by replacing either 250 or 500 g kg− 1 of broken rice from the reference diet. Similarly, two test diets containing lupine were prepared by replacing 75 or 150 g kg− 1 of SBM-56 from the reference diet. Additionally, the pure ingredients were also evaluated against pure broken rice (sorghum, maize or rye) and pure SBM-56 (lupine). In each experiment, the corresponding ingredient was tested at two levels of inclusion and in pure form, using both newly weaned pigs and pigs at four weeks post-weaning. The palatability for each test diet expressed as % of preference was calculated as the percentage contribution of the test diet to total feed intake (test + reference diets). The preferences obtained ranged between 16 and 29% for sorghum, 16 and 35% for maize and 19 and 49% for rye. Except for rye at 500 g kg− 1 (49%), preference values were significantly different from 50% which indicates a higher preference for the broken rice in the reference diet than for sorghum, maize and rye. Preference for lupine ranged between 39 and 56% and no difference in preference was observed between the lupine in the test diets and the SBM-56 in the reference diets. The dietary preferences could already be observed in the first days of experiment, and they did not change substantially thereafter. The levels of inclusion tested did not have an effect on preference for any of the ingredients. The age of the animals did not affect preference and the values obtained in newly weaned and post-weaned pigs were generally in good agreement. However the use of the older animals resulted in higher feed intakes and more accurate measurements. It is concluded that feedstuff palatability can be quantified with a double choice protocol using a reference diet. This should allow the formulation of feeds for weaning pigs taking into account the palatability of the ingredients, in addition to other nutritional criteria. Preference evaluation may become a useful tool to improve the palatability of the diets and facilitate piglets feed initiation at weaning.