By using pulsed eddy-current (PEC), a neural network
　　was adopted to forecast the thickness of metal and
　　the distance
　　between coil and metal. The relationship between the
　　transient current induced and different metallic
　　thickness was analyzed
　　theoretically. The result shows that three features
　　of the signal including the peak height, the time of
　　occurrence of the first peak, and
　　a zero-crossing time, are obtained. The measurement
　　error is within 2%.