As a major greenhouse gas, carbon dioxide (CO₂) plays a critical role in regulating Earth's climate. Small coastal lagoons can act as intense sources of CO₂ to the atmosphere. Emissions from these systems are influenced by multiple factors, including wind speed, water depth, sediment processes, and aquatic biological metabolism. However, CO₂ fluxes from small lagoons remain poorly constrained due to sparse observations and methodological uncertainties. Here, we characterized CO₂ flux dynamics in a small lagoon in South China using three independent approaches, including the floating chamber (FC) method, an air-water boundary layer model (BLM) based on pCO₂ measurements, and an atmospheric gradient transport (GT) model driven by low-cost air sensor observations. Although each method has inherent limitations, including the limited spatial representativeness of FC measurements, uncertainties in air-water exchange coefficient parameterization for the BLM approach, and the need for local eddy diffusivity calibration in the GT model, all three approaches consistently captured the temporal variability of CO₂ exchange after calibration, and the resulting flux estimates showed strong agreement among methods. All three methods identified the lagoon as a net CO₂ source, with mean fluxes of 0.14, 0.20, and 0.18 μmol m⁻² s⁻¹ for FC, BLM, and GT, respectively. The results indicated that CO₂ emissions from the lagoon were primarily driven by heterotrophic metabolism and atmospheric disequilibrium under low-wind conditions. We further evaluated the applicability of UAV-based observations for CO₂ flux estimation. While UAV measurements may help resolve spatial heterogeneity in future studies, they are unlikely to provide reliable flux inversions in highly perturbed nearshore environments. Overall, the strong agreement among the three independent approaches demonstrates the robustness of the observed CO₂ fluxes and highlights the potential of calibrated low-cost sensor-based methods to provide reliable CO₂ flux estimates comparable to those derived from conventional approaches, while enabling cost-effective long-term monitoring.