Five lactating Holstein cows in a 5×5 Latin square experiment were fed five high-concentrate total mixed rations (TMRs) to investigate the effects of step-wise improvement of forage combination on ruminal and milk fatty acid profiles.  The ratio of concentrate to forage was fixed as 61:39, and the step-wise improvement of forage combination was applied as: TMR1, a ration containing corn stover; TMR2, a ration containing corn stover and ensiled corn stover; TMR3, a ration containing ensiled corn stover and Chinese wild ryegrass hay (Leymus chinensis); TMR4, a ration containing the ryegrass hay and whole corn silage; TMR5, a ration containing the ryegrass hay, whole corn silage and alfalfa hay.  The TMRs were offered to the cows twice daily at 0700 and 1900 h.  The entire experiment was completed in five periods, and each period lasted for 18 days.  Diurnal samples of rumen fluids were collected at 0100, 0700, 1300 and 1900 h (day 16); 0300, 0900, 1500 and 2100 h (day 17); and 0500, 1100, 1700 and 2300 h (day 18).  The step-wise improvement of forage combination increased energy and crude protein contents and decreased fibre content.  As a result, the step-wise improvement of forage combination increased dry matter intake and milk yield (P<0.05).  The step-wise improvement increased dietary content of linolenic acid (C18:3n-3), but did not alter dietary proportions of palmitic acid (C16:0), stearic acid (C18:0), oleic acid (C18:1cis-9), linoleic acid (C18:2n-6) and arachidic acid (C20:0).  In response to the forage combination, ruminal concentration of C16:0, C18:2n-6 and C18:3n-3 linearly increased against their dietary intakes (P<0.10).  The step-wise improvement increased milk contents of C10:0, C12:0, C14:0, C16:0, C18:2n-6 and C18:3n-3 (P<0.10) and decreased milk contents of C18:0 and C18:1cis-9 (P<0.05).  Milk yields of C16:0, C18:1cis-9, C18:2n-6 and C18:3n-3 were linearly increased by the increase of these fatty acids in the rumen (R²≥0.79, P<0.05), and milk yields of C18:2n-6 and C18:3n-3 were also positively correlated with dietary intake of these fatty acids (R²≥0.85, P<0.05).  The step-wise improvement increased the transfer efficiencies from feed to milk for C18:2n-6 from 11.8 to 14.2% and for C18:3n-3 from 19.1 to 22.3%.  In a brief, along with the step-wise improvement of forage combination, more dietary linoleic and linolenic acids might escape microbial hydrogenation in the rumen and consequently accumulated in milk fat though these fatty acids were present in low concentrations in ruminal fluids.  The step-wise improvement of forage combinations could be recommended as a dietary strategy to increase the transfer efficiency of linoleic and linolenic acids from feed to milk.