Climate change is intensifying food insecurity across the globe, hitting harder drought-prone rural areas of low-income countries. Yet, limited empirical evidence explains how climate exposure, household adaptive capacity and everyday making of a living jointly shape food security at the local level. This study examines the patterns and determinants of household food security in Guijá district, Mozambique, one of the country’s most climate-stressed regions due to recurrent droughts, floods, erratic rainfall and high temperatures. These climate shocks pose serious threats to agriculture-based livelihoods, food security, water access and overall household welfare particularly because the majority of the population depends on small-scale, rain-fed subsistence farming as their primary economic activity. Using cross-sectional survey data from 407 households, food security was assessed through the Food Consumption Score (FCS), complemented by the Household Dietary Diversity Score (HDDS) and the Household Hunger Scale (HHS). The study applied Multiple Correspondence Analysis (MCA) and K-Means Cluster Analysis to explore the interplay of livelihoods, climate factors and food security outcomes. Descriptive statistics and binary logistic regression identified key determinants of food security, while cluster analysis revealed distinct livelihood patterns. Results indicate that 67.6% of households were food insecure, with notable spatial disparities across communities. Although female-headed households experienced higher levels of food insecurity, gender was not a statistically significant determinant after controlling for socioeconomic and environmental factors. Instead, food security was strongly associated to income levels, income diversification, market access, land-use change and ownership of productive assets particularly motorcycles. Irregular rainfall significantly reduced the likelihood of food security. Four household typologies emerged, with food security probabilities ranging from 0.43 among market-integrated households to 0.09 among wage-dependent households. These findings highlight the need for targeted, livelihood-sensitive and climate-resilient interventions focused on promoting livelihood diversification, improving market and financial access, strengthening climate-smart land management and supporting household asset accumulation for enhanced resilience.