The effects of a SCFP (NutriTek, Diamond V, Cedar Rapids, IA) on feeding behavior and milk production and composition were evaluated in transition cows. Treatments were control (CON; n = 30) or SCFP (n = 34; 18 g/d) TMR fed from −29 ± 5 to 42 d relative to calving. Cows were individually fed 3×/d prepartum (45% NDF, 14% starch) and 2×/d postpartum (32% NDF, 19% starch). Cows were milked 2×/d with samples 2×/wk for composition analysis and calculation of ECM, FCM, and energy balance (EB). Results were analyzed using the MIXED procedure of SAS. SCFP did not affect prepartum DMI, but tended to increase meals/d (8.66 vs. 9.27 ± 0.22 /d; P = 0.06) and reduced time between meals (2.26 vs. 2.09 ± 0.05 h; P = 0.03). A treatment × parity interaction (P = 0.03) reflected greater meals/d in SCFP vs. CON primiparous cows, but not multiparous cows. We observed no difference in prepartum meal weight or length. Similarly, SCFP increased postpartum meal count (11.4 v. 12.6 ± 0.45 /d; P = 0.03) and tended to reduce inter-meal interval (P = 0.07), but did not affect meal weight or length. We observed no differences in milk, FCM, or ECM yields (all P > 0.31). Milk fat concentration was greater in SCFP (P = 0.01) with differences in wk 4 and 5 (P < 0.05) that tended to increase fat yield in these 2 wk, without affecting BCS, BW, or plasma fatty acid concentration (all P > 0.35). Milk lactose concentration tended to be greater for CON (P = 0.06) and milk urea nitrogen tended to be lesser for CON (P = 0.06), but there were no treatment effects on milk protein concentration or somatic cell count. Assuming equal digestibility, EB was lesser in SCFP than CON (−4.34 vs. −6.15 ± 0.74 Mcal/d; P = 0.03) with differences in wk 4 and 5. Energy density of diets calculated from observed ECM yield and BW change did not differ by treatment (1.90 vs. 1.83 ± 0.04 Mcal/kg DM for SCFP and CON; P = 0.18). In conclusion, SCFP increased meals/d and milk fat concentration in wk 4 and 5, although no effect on milk yield was detected.