Invasion of exotic annual grass (EAG), such as cheatgrass (Bromus tectorum), red brome (Bromus rubens), and medusahead (Taeniatherum caput-medusae), could have irreversible degradation impact to arid and semiarid rangeland ecosystems in the western United States. The distribution and abundance of these EAG species are highly influenced by weather variables such as temperature and precipitation. We set out to develop a machine learning modelling approach using a lightGBM algorithm to predict how changes in annual and immediate past precipitation regimes impact the abundance of EAG in the study area. The predictive model primarily utilized edaphic and weather variables and a seed source proxy from previous years to make the predictions. We achieved strong training accuracy (r= 0.95 and MdAE=2.36 of percent cover) and test accuracy (r= 0.79 and MdAE=4.54 of percent cover). We predicted five versions of EAG percent cover maps for 2022 with different precipitation scenarios, i.e., with the 9-year average, half of the average, three fourth of the average, one and half of the average, and twice the average precipitation. Five versions of spatially explicit EAG percent cover 2022 datasets can provide valuable information to local and regional land managers so they would know what EAG abundance would look like with certain precipitation scenario.