The areas under turfgrass increase with the increase in rapid urbanization rate. Therefore, accounting for their contribution to greenhouse gas emission to atmosphere is needed. Information in this area is rather limited. This study aimed to quantify and model two years long term (2016-2017) temporal variations in soil CO2 emissions from newly established turfgrass area under semiarid conditions and to assess the influence of irrigation on the soil CO2 respiration. The measurements were obtained from irrigated and non-irrigated plots with respect to meteorological variables (air temperature, soil temperatures at different depths, rainfall and relative humidity) and the field soil moisture data using classical and Stepwise MLR models. Potential use of available auxiliary meteorological and field soil moisture data for the estimation of future CO2 emissions were also investigated using Artificial Neural Network (ANN) models whose accuracy was tested using independent data sets. Two years of continuous measurements revealed that CO2 emissions highly varied seasonally from 318 to 2621 kg ha(-1) week(-1) for irrigated and from 390 to 1711 kg(-1) ha(-1) week(-1) for non-irrigated plots, respectively, with 31% more emissions from irrigated plots compared to non-irrigated ones. Inclusion of the soil field moisture data did not significantly impact the results. ANN models were able to moderately successfully estimate CO2 emissions.