The transition from winter-feeding to pasture involves a considerable diet change that might compromise health of horses, and info on the metabolic responses in horses on pasture are scarce. The objective of the study was to investigate how the transition from haylage feeding to strip grazing affected feed intake (FI) and metabolic responses (blood glucose (GLU)-insulin (INS) and cecal pH). Four cecum-cannulated Norwegian cold-blooded trotter geldings were used in a longitudinal study consisting of 3 periods: 1) haylage only (d 0 to 8) reflecting winter-feeding, 2) transition (d 9 to 20) reflecting a diet where the proportion of forage decreased and grazing increased (0 min to 2 × 3 h per d), and 3) grazing only (d 21 to 28) reflecting horses on pasture (2 × 3 h per d). Strip grazing was used to control and measure daily FI in period 2 and 3. From d 9 to 28 the horses were allowed access to a new strip of grass for a specified time-period. The remaining grass was cut each day after grazing, collected, weighed, and compared with the amount of grass from a strip of ungrazed pasture. It was possible to estimate FI for all horses and calculate an average FI per horse. Haylage and grass samples from every 4th d were analyzed for nutritional composition (neutral detergent fiber excluding residual ash (aNDFom), water soluble carbohydrates (WSC), and crude protein (CP)) and used to estimate dry matter (DM) and nutrientintake. On d 8 and 28, blood and cecal samples were collected before and each h after feeding (2 kg haylage as fed vs. 3 h of grazing) for 8 h. ?Area under the curve (AUC) of GLU and INS were calculated using GraphPad Prism. The effect of d on nutrient content, FI, body weight (BW), AUC and peak GLU and INS, and minimum pH were compared using PROC mixed in SAS and considered different if P < 0.05. Grass yield increased (P < 0.001) over time as the grass matured (0.32 to 0.54 kg DM/m2). Plant maturity of the grass affected the nutrient content (P < 0.001) as aNDFom increased (545 to 636 g/kg DM), and CP decreased (158 to 101 g/kg DM). The WSC content varied (P < 0.001) over time (73 to 128 g/kg DM). Similar BW was measured on d 8 and 28 (568 ± 31 kg). The DM intake was 1.62 kg (0.29% of BW) on d 8 and estimated to be 3.7 kg (0.65% of BW) on d 28. There was no effect of d on AUC and peak GLU and INS, but cecal pH was higher (P < 0.001) when feeding haylage (6.89) on d 8 than pasture (6.57) on d 28. In conclusion, FI was controlled by strip grazing, BW was maintained, and metabolic responses were within the expected ranges. However, a longer grazing time would probably result in increased FI and undesired BW gain.