Meteorological drought and armed conflict may jointly shape acute food insecurity, yet their conditional and spatial associations remain insufficiently understood in conflict-affected settings. We examined whether prior local conflict exposure modified the association between drought severity and numerically coded food-insecurity phases in Tigray, Ethiopia, and whether related interactions extended across neighbouring woredas. We constructed a monthly panel of 93 administrative districts (woredas) for January 2010–May 2024 (16,089 woreda-months), combining Integrated Food Security Phase Classification (IPC), CHIRPS v3 derived Standardised Precipitation Index (SPI-3), preceding six-month conflict exposure recorded by the ACLED, and vegetation deficit. Woreda and year-month fixed effects were incorporated into spatial econometric models, with a Queen-contiguity Spatial Durbin Error Model (SDEM) as the preferred models. We assessed robustness using alternative spatial models, marginal-effects analysis, and residual diagnostics. Prior local conflict positively moderated the association between drought severity and numerically coded food-insecurity phases (interaction β = 0.0197, 95% CI [0.0082, 0.0312]). The interaction remained positive across six specifications, although its magnitude varied. Conditional marginal-effects analysis showed that a clearly positive drought association emerged only at unusually high conflict exposure, where observations were sparse. Substantial spatial dependence was identified in unobserved disturbances, whereas directional neighbouring drought–conflict interactions were weak, imprecise, or specification-sensitive. These observational associations do not establish causal effects or spatial transmission mechanisms. The findings suggest that integrating recent conflict exposure and spatial information into drought-based monitoring may improve the contextual interpretation of acute food-insecurity risks, particularly in areas experiencing intense conflict.