Dairy development provides substantial economic opportunities for smallholder crop-livestock farmers in East Africa, but productivity is constrained by scarcity of quantity and quality feed. Livestock is also associated with negative environmental impacts, including greenhouse gas (GHG) emissions, air pollution, high water consumption, loss of biodiversity and land degradation. Improved livestock feeding has been proposed as a triple-win strategy towards achieving climate-smart agriculture, increasing food security and climate change adaptation, and decreasing GHG intensities. Improved tropical forages include a wide variety of sown or planted annual and perennial grasses, herbaceous or dual-purpose legumes and shrubs developed for increasing productivity of grazed and cut-and-carry fed livestock. This thesis aims to explore potential impacts and trade-offs associated with the implementation of improved livestock feeding and forage technologies at farm scale, across a diversity of smallholder crop-livestock systems in East Africa. We first quantitatively reviewed 73 published studies to take stock of evidence on agronomic, livestock, environmental and economic impacts of tropical forage technologies in Sub-Saharan Africa. We then introduced a relatively simple approach to quantify feeding systems and feed gaps in data-scarce smallholder systems. Based on this, household-level impacts and trade-offs of improved livestock feeding and forages were explored in Tanzania and Rwanda, considering productivity, environmental and livelihood dimensions. To do so, a combination of approaches and methods was employed, including coupled livestock and economic modeling, bio-economic farming systems modeling and multi-objective optimization, multi-variate statistics, on-farm monitoring and measurements, semi-structured interviews, participatory validation and expert knowledge. Results from this thesis confirmed the considerable feed gaps in metabolizable energy and crude protein. Feed gap here is defined as the difference between livestock feed demand for an attainable milk production level (attainable feed demand) and actual feed supply at individual herd level. The literature review revealed that higher herbage production and quality of improved forage technologies resulted in an average 35% milk and 24% manure increase and 60% higher associated food crop yields while almost halving soil loss. In Lushoto in Tanzania, predicted adoption rates for improved livestock feeding and poverty reduction among households with improved dairy cows were likely to be higher compared to households with only local cows. Methane emissions intensity declined with adoption of improved livestock feeding strategies. Across Rwanda, livestock production was among the most important pathways to higher food availability. Total baseline GHG emissions were low, with enteric fermentation and manure emissions contributing 74 – 81%. Scenario assessment of three policy options showed that Girinka program (providing a crossbred cow to a poor households) strongly increased food availability of the poorest section of the household population at a high GHG cost, while improved livestock feeding had less impact on food availability but at an only small GHG increase. In Babati in Tanzania, emission intensity was lowest for the dairy farming system (2.1 kg CO2e kg-1 milk) when compared to three other smallholder livestock types, with lowest trade-offs with household income, and carbon and nitrogen balances. Available options to reduce agro-environmental trade-offs included reducing ruminant numbers, replacing local cattle with improved dairy breeds, improve feeding through on-farm forage cultivation to reach higher milk production levels, and reducing crop residue feeding to leave them on the field. However, adoption of these technologies require a skillful re-organization of the entire production system, resulting in loss of some multi-functionality of livestock, and incur higher production risks. In conclusion, we argue that pathways to sustainable intensification of livestock in East Africa are needed, which can also unlock new financing avenues. Improved forages can play a key role as they deliver multiple productivity, economic and environmental benefits when skillfully integrated in smallholder cropping and farming systems. Livestock feeding and forages are thus at multiple crossroads: at a point where crucial decisions regarding future pathways are taken, where productivity and environmental impacts meet, and where scientific disciplines including agronomy, soil and animal science intersect. Results from this thesis aim to inform policy makers, project designers, investors, donors and other decision-makers on prioritizing options towards low emission livestock development. Suggestions and recommendations for future research include next-level forage agronomy, further multi-disciplinary and systems-level trade-off analysis, and quantification of technology contributions to national-level climate and land restoration policy targets.