AbstractIn the network‐based on‐site earthquake early warning system (EEWS), the ‘blind zone’, namely the zone where the issued warning arrives later than the destructive S and surface waves, is one of the challenges affecting its effectiveness. The ‘blind zone’ is determined by the interstation distance, or equivalently the density of seismic stations, of the network. In this paper, we suggest a practical approach according to which, when in a region a temporary increase of seismic hazard is declared, additional stations are deployed in such a way that the blind zone is temporarily reduced. In the procedure, the time‐dependent neo‐deterministic seismic hazard assessment (TD‐NDSHA) plays a vital role in the identification of the regions potentially exposed to high macroseismic intensities. As a showcase example, we consider the scenario of year 2014 at the Sichuan‐Yunnan border of southwest China. The TD‐NDSHA is based on the standard NDSHA procedure at regional scale (bedrock conditions), with the ‘controlling earthquakes’ defined on the basis of the Annual Consultation. We show that the ‘blind zone’ can be reduced in the identified areas of interest (e.g., MMI ≥ VI), by deploying a limited number of additional seismic stations. In the case where false alarms can be tolerated, significant reduction of the ‘blind zone’ can be implemented by moving from a network‐based EEWS to a single‐sensor‐based EEWS and skipping the process of location and magnitude‐determination/prediction procedures.