In vitro fermentation of current forage diets under tropical agroforestry systems were evaluated. Two common tropical grasses (Panicum maximum cv. likoni, PM; Pennisetum purpureum cv. CT-115, PP) were supplemented with four tree/shrub forage (Leucaena leucocephala, LL; Gliricidia sepium, GS; Trichantera gigantea, TG; Morus alba, MA). Representative forage samples were randomly selected in an agroforestry experimental eld of Matanzas, Cuba (22°482N, 81°22W; 60 masl). The in vitro gas production (IVGP) technique of Theodorou et al. (1994) was used in two consecutive incubation series (96h), using rumen inoculum from two ruminally cannulated Holstein Frisian dairy cows, previously adapted. A factorial arrangement design (2×4) was used to evaluate combination of grasses (n = 2) and trees or shrubs forage (n = 4) as supplements (ratio 60:40 of grass:supplement). Treatments were: 1) PM-LL; 2) PM-GS; 3) PM-TG; 4) PM-MA; 5) PP-LL; 6) PP-GS; 7) PP-TG; 8) PP-MA. Fermentation pro les of grasses (PP; PM; without supplement) were considered as 3controls3. The IVGP, fermentative parameters (a; b; c; 3lag time3, T), and in vitro disappearance of dry matter (IVDDM) and neutral detergent ber (IVDNDF) were determined. Gas production (GP) data were tted to the France et al. (1993) equation. There were signi cant differences (P< 0.05) in fermentative patterns among diets. The IVGP (PP > PM), and IVDDM or IVDNDF, were higher and lower (P< 0.05), respectively, for sole grasses compared to their respective combined diets. Except diet PM-GS (highest IVGP values), IVGP was lower (P< 0.05) in supplemented diets when compared to control (PM) despite its T value (2.18), degradation rates (μ) or IVDMD (559.4 g kg-1DM). Supplement TG resulted with the lowest fermentative values for both grasses.