Agroforestry is being promoted as an alternative to sole cropping in semi-arid regions, yet farmers lack evidence-based clarity on how different planting patterns influence productivity, carbon stock, biodiversity, and economic viability. This study addresses this gap by evaluating agroforestry performance in Jagtial and Peddapalli districts of Telangana, India. Thirty villages (19 in Jagtial and 11 in Peddapalli) were sampled, selecting farmers across marginal, small, semi-medium, and medium landholding categories. Eight planting patterns were identified, with bund planting being most common (37.42%), though block + boundary had the highest tree density (154.22 ha-1 in Jagtial, 140.70 ha-1 in Peddapalli). Bund planting had the lowest density but highest aboveground biomass per tree in Jagtial (291.61 kg tree-1), while boundary planting led in Peddapalli (257.11 kg tree-1). Sole paddy yield was 5.10 t ha-1 in Jagtial and 4.67 t ha-1 in Peddapalli. Block + boundary yielded the highest biomass and carbon stock (32.98 t ha-1 and 15.50 t ha-1 in Jagtial; 22.26 t ha-1 and 10.46 t ha-1 in Peddapalli). Jagtial outperformed Peddapalli in total biomass (16.35 vs. 11.74 t ha-1) and carbon stock (7.68 vs. 5.64 t ha-1). Species richness was greater in Jagtial (46 species) than in Peddapalli (35), with Mangifera indica dominant. Block + boundary had the highest B:C ratio (3.67 in Jagtial, 3.47 in Peddapalli). The findings provide practical guidance for farmers and policymakers seeking climate-smart, profitable, and biodiversity-enhancing land use options.