<![CDATA[Human health concerns are one of the important consequences of low air quality. The low air quality of each city will have long-term impacts such as global warming and anthropogenic greenhouse effects. Air quality usually occurs in areas that are in some parts of the country such as Kalimantan Island. As the third largest island in the world, Kalimantan can be said to be the lungs of the world like the haze problem that enveloped the city of Banjarmasin in 2019. This condition can result in high sufferers of Acute Respiratory Tract Infection (ISPA). Decision making by stakeholders needs to be studied in depth to prevent this. One of the efforts that can be done is the air quality forecast that will occur. Data obtained from BMKG Banjarmasin is the initial material for the forecast. Air quality forecast will use Triple Exponential Smoothing with 2 types of modeling namely additive and multiplicative, so this study aims to conduct air quality forecasts in Banjarmasin City in 2021 and 2022 using Additive and Multiplicative Triple Exponential Smoothing. In forecasts using this method, weighting the constant values α, β, γ can result in small error values. To determine the accuracy comparison of the two modeling is done with an RMSE value. The results showed that air quality conditions in Banjarmasin during 2021 and 2022 for CO, O3, and PM pollutants were in the category of safe for human health, while for pollutants NO2 and SO2 were declared to have a high index so that air quality can harm the health of living things. In comparison, multiplicative modeling on CO forecasts (α= 0.5, β = 0.001, and γ = 0.149), NO2 (α = 0.5, β = 0.024, and γ = 0.022), and SO2 (α = 0.5, β = 0.001, and γ = 0.037) has high accuracy and small error values compared to additive modeling. In contrast, additive modeling in O3 (α = 0.5, β = 0.001, and γ = 0.06) and PM (α = 0.434, β = 0.001, and γ = 0.213) have high accuracy and low error values compared to multiplicative modeling. The conclusion obtained is the difference in forecast results between additive and multiplicative modeling on air quality forecasts in Banjarmasin because multiplicative modeling is used when there is a trend or sign that seasonal patterns depend on the size of the data. In other words, seasonal patterns enlarge as the data size increases. Additive models are used if this trend does not occur.]]>
