Accurate quantification of greenhouse gas (GHG) emissions from agriculture is critical for establishing credible baselines, identifying emission hotspots, and prioritizing mitigation strategies. In Kenya, most agricultural emission estimates have relied on nationally aggregated activity data and Tier 1 emission factors, which limit the representation of spatial heterogeneity and management effects. This study employed a spatially explicit bottom-up framework that combines IPCC default and default-derived parameters and emission factors, empirical model-derived approaches, and regionally differentiated livestock parameters, integrating subnational production data, management practices, and emission intensities to generate county-level estimates for major crop and livestock systems. In 2021, total agricultural emissions were estimated at 26.74 Mt CO2e, with livestock accounting for 22.90 Mt CO2e (86%) and crops accounting for 3.83 Mt CO2e (14%). Maize represented the largest share of crop emissions (38.9%), while rice exhibited the highest emissions per unit area (3.6 t CO2e ha−1). Emissions were spatially concentrated, with a small number of counties responsible for a disproportionate share of the national total. Scenario analysis revealed substantial mitigation potential in both sectors. For maize, optimized nitrogen management reduced emissions by 100–140 kg CO2e ha−1, and integrated strategies achieved reductions exceeding 200 kg CO2e ha−1; in rice the corresponding reductions were 279 and 303 kg CO2e ha−1. In cattle systems, to which the livestock mitigation analysis was restricted, fodder-based and residue-based feeding strategies, as well as anaerobic digestion, demonstrated the highest mitigation potentials, with reductions varying across grazing, mixed, and stall-fed systems. Qualitative marginal abatement comparisons suggest that management-based interventions are relatively low-cost and scalable, while integrated or technology-intensive measures achieve greater reductions but require stronger enabling conditions. These results show that spatially explicit bottom-up accounting improves upon nationally aggregated Tier 1 inventories without constituting a complete Tier 2 assessment and provides a stronger basis for prioritizing mitigation pathways.