Abstract: The current identification system mostly USES the password identification technology, easy to be stolen or lost, leading to loss of users have a certain, the identification accuracy is low.To this end, we design a new network identification system based on web, the designed system is mainly composed of browser side recording module, voice print data transmission module and voiceprint training and recognition modules, system architecture is given.Based on the system architecture in the browser to install an Activex controls, recording on the user's voice recording.Through the pre-emphasis, endpoint detection methods such as the voice print signal preprocessing.Characteristic value of voiceprint signals are extracted through the signal conversion to filter the useless part of the signal, to extract the characteristic parameters of essential attributes can reflect user recording voice, according to the feature extraction is worth to D value, through the D value identification.The experimental results show that the designed system has a high recognition accuracy.