Anchorage, Alaska, is a natural laboratory for recording strong ground motions from a variety of earthquake sources. The city is situated in a tectonic region that includes the interface and intraslab earthquakes related to the subducting Pacific plate and crustal earthquakes from the upper North American plate. The generalized inversion technique was used with a local rock reference station to develop site response at strong‐motion stations in Anchorage. A database of 94 events recorded at these sites from 2005 to 2019 was also compiled and processed to compare their site response with those in the 2018 7.1 event (main event). The database is divided into three datasets, including 75 events prior to the main event, the main event, and 19 aftershocks. The stations were subdivided into the site classes defined in the National Earthquake Hazards Reduction Program based on estimated average shear‐wave velocity in of the upper 30 m (), and site‐response results from the datasets were compared. Nonlinear site response was observed at class D and DE sites ( of 215–300 and , respectively) but not at class CD and C sites ( of 300–440 and , respectively). The relationship of peak ground acceleration versus peak ground velocity divided by (shear‐strain proxy) was shown to further support the observation that sites with lower experienced nonlinear site response.