Bangladesh combines one of the world's highest agricultural population densities with acute exposure to flooding, drought, salinity intrusion and cyclonic surge, making the country a critical test case for whether sustainable agricultural practices (SAPs) can deliver climate resilience and rural economic development simultaneously. This paper reports empirical findings from a ten-year (2016–2025) balanced household panel of 3,100 farm households drawn from eight districts spanning six agro-ecological zones, complemented by 1,240 georeferenced soil cores, 96 focus group discussions and 48 key-informant interviews. Eight practices were assessed: conservation tillage, alternate wetting and drying (AWD), crop diversification, integrated nutrient management (INM), agroforestry, cover cropping, integrated pest management (IPM) and on-farm water harvesting. Treatment effects were identified using propensity-score matching, inverse-probability-weighted regression adjustment, endogenous switching regression and household fixed-effects panel models, with Rosenbaum sensitivity bounds and placebo tests for robustness. Adoption of at least three practices raised rice-equivalent yield by 0.71 t ha⁻¹ (12.7%), net farm income by BDT 21,400 ha⁻¹ (21.0%) and the multidimensional resilience capacity index by 14.6 points, while reducing irrigation water application by 284 m³ ha⁻¹, greenhouse gas intensity of paddy by 0.31 kg CO₂e kg⁻¹ (22.0%) and the household food insecurity access scale by 2.1 points (all p < 0.01). Soil organic carbon under conservation agriculture rose from 1.02% to 1.57% over seven years against a declining conventional baseline. Village-level income multipliers averaged 2.63, female labour participation increased by 42%, and seasonal out-migration fell by 34%. Benefit–cost ratios ranged from 1.81 to 4.00 at a 10% discount rate. Adoption was constrained principally by upfront cost (64% of households), knowledge gaps (58%) and labour scarcity (47%); extension access, farmer-group membership and credit availability were the strongest positive determinants. Scenario analysis indicates that scaling SAPs to 75% of cultivated area by 2040 would raise the national food production index to 149 (2025 = 100) while lowering agricultural emissions to 57. The findings support a bundled policy package combining concessional green refinance, index-based insurance, decentralised agro-service hubs and measurable soil-carbon monitoring.