BACKGROUND: Mozambique has one of the lowest electrification rates in the world, leaving its population dependent on polluting fuels for lighting. Limited epidemiological evidence links fuel-based lighting to child health. We examined associations between polluting lighting sources, particularly kerosene, and under-five morbidity in a semi-rural district in Mozambique. METHODS: We constructed a birth cohort from demographic and hospital-based pediatric morbidity surveillance data. We included children born in Manhiça district between January 1, 2016, and December 31, 2020. The most common lighting source (polluting vs clean) used during follow-up was used as exposure. The outcome was the frequency of all-cause and respiratory-linked hospital visits in any of the seven surveilled health facilities within the district. We used zero-inflated negative binomial regression models to calculate Rate Ratios (RR) adjusted for potential confounders at the child, mother, and household level. RESULTS: We included 17,815 under-five children (49% female) living in 13,574 households. Nearly a quarter of children (24.7%) lived in households using polluting lighting fuels. During follow-up, there were a total of 69,677 all-cause hospital visits (53% respiratory-linked). Children in households with polluting lighting had a 2% higher rate of all-cause hospital visits (RR: 1.02, 95% CI: 0.98, 1.06) and a 2% lower rate of respiratory-linked hospital visits (RR: 0.98, 95% CI: 0.94, 1.03) compared to those with clean lighting. Results were robust to sensitivity analyses testing alternative inclusion criteria based on the definition of health facility catchment area and duration of follow-up time covered by the surveillance system. CONCLUSION: Polluting compared to cleaner lighting sources were not associated with hospital visits among children under five. Intervention-based research is needed to better understand the health impacts of air pollution from fuel-based lighting among children living in areas with limited access to electricity.