The present study was conducted to evaluate the characteristics of electroacupuncture (EA)-induced analgesia in mice. Three inbred strains of mice (DBA/2, C57BL/6J, BALB/c) and three outbred strains (ICR, LACA, NIH) were used in the experiment. Two pairs of metallic needles were inserted into acupoints ST 36 and SP 6 connected to an electric pulse generator. EA parameters were set as constant current output with alteration of a positive and negative square wave, 0.6ms in pulse width for 2Hz and 0.3ms for 100Hz. Tail-flick latencies evoked by radiant heat were measured before, during and after EA stimulation. We found that (1) DBA/2 mice showed a significantly more potent analgesic effect than the other five strains in response to both 100 and 2Hz EA. In this case, the intensities were 1.0-2.0-2.0mA, 10 min for each intensity totally 30min. (2) EA analgesia increased as the intensity of stimulation increased from 0.5 to 21 2.0mA, but it remained at this plateau when the intensity further increased from 2.0 to 3.0mA. (3) 10.0mg?kg naloxone was needed to 21 block the analgesic effect induced by 2Hz EA of 2.0mA, but to block that by 100Hz, 25.0mg?kg was necessary. (4) A positive 21 correlation was observed between analgesia induced by morphine at the dose of 5.0mg?kg and by 100Hz EA in two tested strains DBA/2 and C57BL/6J. In conclusion, EA induces reliable, strain-dependent analgesia in mice. The naloxone-reversibility of EA, a measure of whether it is opioid or non-opioid mediated, is dependent upon intensity and frequency.