The gold standard method for measuring digestibility of feedstuffs is by total collection of feces. An alternative method is the in sacco technique which provide information on both digestibility and digestion kinetics of a feed. The in sacco technique is commonly used in feedstuff evaluation systems for ruminants. Only few studies have used the technique in horses, but further validation against in vivo studies is needed. The objective of the study was to determine if the in sacco technique can be used to measure digestibility in horses. Five cecum cannulated Norwegian coldblooded trotters (body weight 550 ± 75 kg) were used in a 22 d study consisting of 18 d of diet adaptation followed by 4 d of total collection of feces. Horses were fed 24.3 kg of a late harvested silage (LHS, dry matter (DM) 31.0%, neutral detergent fiber (NDF) 59.6% of DM) fed in 3 equal meals at 0800, 1400 and 2000 h. The apparent total-tract digestibility (ATTD) of LHS was estimated by total collection of feces over 4 consecutive days. Before the collection period, in sacco digestibility was estimated of the LHS and an early harvested silage (EHS, DM 26.0%, NDF 28.6% of DM) using the in sacco technique (bag size 6 × 15 cm, 5.13 mg feed/cm 2 , pore size of 36 μm). The EHS and LHS were dried, ground to 1 mm and 1 g was added to in sacco bags. Two bags of each feedstuff were fastened to a 78 cm long string with rubber bands and incubated in the cecum for 0, 2, 4, 8, 16 and 24 h, then washed, dried for 48 h and weighed to determine DM loss after incubation. The in sacco digestibility was modeled in SAS according to D = A tB (e −Ct ) , where D is the digestibility through time t, A is the soluble fraction of the feedstuff and curve interception, B is the potential digestibility of the insoluble fraction, and C is the digestion rate of B. The potential digestibility (A+B) of the LHS was compared with EHS and to the ATTD of LHS in SAS using a paired t -test and considered significant if P < 0.05. The ATTD of DM for LHS (57.0% ± 0.49) was not different ( P = 0.22) from the potential digestibility (58.4% ± 2.16) measured with the in sacco technique. The A, B and C values were 33.6, 24.8 and 0.096 for the LHS and 64.2, 28.6 and 0.178 for the EHS. The potential digestibility for EHS (92.8% ± 0.42) was higher ( P < 0.001) than the LHS as expected and show that the technique can differentiate between feedstuffs. In conclusion, the in sacco technique resulted in similar results as in vivo ATTD of LHS and it is a promising technique for estimating digestibility of individual feedstuffs in horses. However, the technique needs to be developed further and validated using other feedstuffs and mixed diets.