Air pollution is a serious threat to public health worldwide, causing millions of premature deaths each year. In Latin America, and especially in Ecuador, the situation becomes even more complicated due to poor monitoring infrastructure, making it difficult to effectively track pollutants such as PM₂.₅, CO₂ and CO. This research presents an innovative approach that is based on the calibration and validation of low-cost sensors (ME) to measure pollutants in outdoor spaces, using multivariate mathematical models (A, B, C and D) that include temperature and relative humidity as additional variables. Platforms with ME1 and ME2 sensors were used to collect PM₂.₅, CO₂ and CO data for more than two weeks, comparing these data with reference instruments. The calibration models were evaluated using metrics such as MAE, R², MAPE and SMAPE. The results indicated that, once calibrated, the ME1 and ME2 sensors achieved correlations above 92% with the reference instruments for all pollutants, with absolute and relative errors within acceptable ranges. The inclusion of environmental variables consistently improved the fit of the models, especially model D. Among the four approaches evaluated (A–D), Model D was the most efficient for PM₂. ₅, reaching R² = 0.963 and MAE = 3.24 µg/m³ in ME1 (8.5% improvement vs. Model A), and R² = 0.957 and MAE = 3.92 µg/m³ in ME2 (8.2% improvement). For CO₂, the differences between models were small (ME1: max. R² = 0.839 in D; ME2: max. R² = 0.763 in D), with MAE variations < 1%. For CO, the best performance was marginal and depended on the metric: in ME1, the lowest MAE was obtained by Model B (54.83 ppb), while in ME2, Model D (57.48 ppb) reduced the error by ~1.2% compared to A, with R² ≈ 0.929. This study demonstrates that, with proper calibration, low-cost sensors can be effective tools for air quality monitoring in resource-limited contexts such as Ecuador, strengthening environmental surveillance strategies and facilitating evidence-based decision making.