Introduction: The poultry industry involves billions of animals and, thus, constitutes a priority for animal welfare. Objective: To compare closed-sided (CS) and open-sided (OS) industrial houses regarding broiler chicken welfare in Southern Brazil. Methodology: Ten flocks in each house type were evaluated. Relevant measures were organised into four groups: 1) health indicators: contact dermatitis on the breast and abdominal areas (CDE), bird soiling (BSO), foot pad dermatitis (FPD), hock burn (HBU), lameness (LAM), fractures (FRA), bruising (BRU), scratches (SCR), dead on arrival (DOA), diseases (DIS); 2) environmental indicators: relative humidity (RHU), temperature (TEM), air velocity (AVE), ammonia (NH3), carbon dioxide (CO2), light intensity (LIN), litter moisture (LMO); 3) behavioral indicators: bird behavior (BBE), touch test (TTE); 4) bird affective states: qualitative behavior assessment (QBA). Generalized linear models were fitted to CDE, BSO, FPD, HBU, LAM, DOA, FRA, BRU, SCR, DIS; linear models to RHU, TEM, AVE, NH3, CO2, LIN; t-test to LMO; regression model to count data and generalized linear models to BBE; Principal Component Analysis for QBA. Results: The welfare indicators for CS as compared to OS houses presented worse values for CDE (P=0.026), exploratory behavior prevalence (P=0.012), NH3 (11.2±6.8 vs. 7.5±3.9 ppm) and CO2 concentrations (1124.9±561.5 vs. 841.0±158.0 ppm), and LIN (6.9±6.3 vs. 274.2±241.9 lx). On the other hand, OS house presented higher prevalence for SCR (P=0.043) and panting behavior (P=0.002), and lower AVE (2.1±0.7 vs. 1.1±1.0m/s). Stocking density (13.9±0.4 for CS vs. 12.0±0.3 birds/m2 for OS houses) likely influenced some results. Discussion: Results for welfare indicators per house type were mixed, as CS houses presented better results for SCR, panting behavior and AVE; and OS houses presented better CDE, exploratory behavior, LIN and gas concentrations, probably due to curtain handling. The birds explored the environment more in OS than CS houses, perhaps due to higher LIN and lower stocking density; however, these birds faced more heat stress, as shown by higher panting behavior, perhaps due to lower AVE in OS than CS houses. Conclusion: Open-sided houses presented better results for five indicators as compared to three for CS houses; overall, both house types presented important welfare problems.