Abstarct: Abstract The present research is concerned with modeling natural gas consumption and its prediction in the domestic sector of Golestan Province. The studied statistical population includes the urban subscribers of 14 cities of the province in the household sector in three units: city, urban household and citizen, and the study period covers the years 2014-2022 (96 months). The panel data regression model and R software were used in modeling and forecasting. According to the results, weather variables, income and consumption of the previous period had a greater influence on the urban consumption model. In addition, in the household consumption model, the weather variables, the number of holidays, and the household's consumption before and after the period have a greater influence. In the citizen consumption model, weather variables, the number of holidays, the consumption of the previous period and the gas access index have a greater influence. Furthermore, the forecasts of the three models studied are 24 future periods in the city and citizen consumption models and 1 future period for the urban household consumption model.