The discovery of a new exotic insect herbivore triggers responses from biosecurity agencies, one of which is the decision of whether or not to attempt eradication. Rapid determination of the host range of the new invader is necessary, but when sap-sucking insects are first collected from plants, the lack of visible signs of feeding damage makes it difficult to determine their host status. We investigated the Electrical Penetration Graph (EPG-DC) technique as tool to assess host range of a xylem sap-feeding invader, using Carystoterpa fingens (Hemiptera: Aphrophoridae), a New Zealand endemic xylem feeder, as a model insect. Real-time probing and feeding events over a 12-h recording period were compared for adult C. fingens on 18 plant species. Hebe azure, a known host, was designated the ‘reference plant species’ against which events on all other plants were statistically compared. EPG waveforms were categorized on their amplitude, frequency, voltage and electrical origin, and six parameters (time taken to first probe, time to first xylem ingestion from first probe, total probing time, number of xylem-ingesting events, duration of the longest xylem-ingesting event and total xylem ingestion time) were measured. The total xylem ingestion period (i.e. the actual feeding period) on each plant species expressed as a percentage of total probing time was considered the best parameter for comparing the host status of plants with H. azure. Although the EPG data overestimated the actual host range of C. fingens, we consider that they provided a reasonable first guide to the potential host status of the unknown plants, and so might usefully be used to rapidly assess whether a plant from which a new invader was collected was a host, or whether the association was merely incidental.