The present thesis aimed to study the ensilage complete feed production from agricultural by-products for dairy cattle in Thailand. This study comprised 4 experiments. The first experiment detennined chemical composition and dry matter degradability of various agricultural by-products. The results showed values within the same range generally reported. The second experiment was conducted to investigate the chemical composition and degradability of various ensilage complete feeds with varying ensilage time. The experimental design was 5 x 3 factorial arrangement in CRD, which factor A was level of urea addition and factor B was time of ensilage. Chemical composition changed little with time and slightly varied among levels of urea. DM degradability increased with increasing cassava level while CP degradability and pH level increased with increasing urea addition. By using Flieg scoring which related to VFA yields, there were no significant difference among ensilage complete feeds and times of ensilage. Therefore, it can be concluded, in this experiment, that the 5(th) ensilage complete feed is more appropriate since its DM and CP degradability were highest. The third experiment was carried out to determine the quality of the 5(th) ensilage complete feeds (Exp. 2) after being storage for 6 months. The experimental design was a CRD arrangement. Samples were taken at 1 month interval up to 6 months and were subjected to laboratory and degradability analyses. The results showed no significant (P>0.05) difference in chemical composition except for increased NDF and ADF percentage in association with increasing time of storage. By using Flieg scoring which related to VFA yields, there were no significant (P>0.05) difference among times of storage. In conclusion, this experiment showed that the ensilage complete feed can be stored for more then 6 months. The final experiment was conducted to investigate the effect of ensilage complete feed on performances of dairy cow in early lactation. Sixteen Holstein-Friesian crossbred lactating cows, with averaging 14.5(+,ฑ)3.6 kg milk/day, 73(+,ฑ)28 days in milk and 420(+,ฑ)52 kg liveweight, were stratified random balanced into two groups (8 cows each group). The first group was fed meal concentrate with grass silage while the second group was fed ensilage complete feed. The cows in the first group consumed more DM and ME than those cows in the second group. Milk yields and milk composition were also higher in group 1 than in group 2. It can be concluded in the present study that the ensilage complete feed was not appropriate for feeding to the lactating cows. However, before further conclusion will be made, more researches are needed particularly on the method of producing the large scale ensilage complete feed.