In order to improve the automatic control effect and speed up the translation speed, an automatic control system for translation robots based on intelligent speech is designed. Collect external intelligent voice signals, obtain digital signals using an A/D converter, activate the voice wake-up module to activate the translation robot, recognize complex voice signals through dictation mode, recognize simple voice signals through command mode, and obtain language and text recognition results. Extract Keywordsas automatic control instructions for the translation robot through a deep learning Keyworddetection method, and identify automatic control instructions through a microcontroller. Experimental results show that the system can effectively collect external intelligent voice signals, extract Keywordsfrom intelligent voice signals, and complete automatic control of translation robots.