Voice recognition system capable of loading personalized feature model, and method

The invention provides a voice recognition system capable of loading a personalized feature model, and a method. The system comprises a voice coding chip which is used for carrying out the A/D conversion of a received analog voice signal to obtain a digital audio signal, and carrying out the D/A conversion of the digital audio signal processed by a digital signal processor into an analog voice signal; the digital signal processor which is used for carrying out the voice recognition algorithm processing of the inputted digital audio signal; a FLASH chip which is used for storing a voice recognition program of the digital signal processor and universal voice model data, wherein the program and voice model data are loaded to a DDR RAM chip from the FLASH chip after power-on start; and the DDR RAM chip which is used for operating the voice recognition program, and stores the universal voice model data and personalized feature model data. The system can be used for the UI voice control technology, can load the per