The paper, based on the concept of FOOTING, makes a multimodal discourse analysis of the relationship between the judge’s discourse and his footing shifts in a criminal courtroom. The results show that in the interaction, multimodal resources in judges’ discourse include conversational features (prolonging keywords, interrupting, repeating, taking turns, etc.), acoustic ones (ascending F0 for pitches and dB for intensity, transition tracks between consonants and formants of vowels, duration of some keywords in important sentences, etc.), and visual ones (facing other parties, facing the materials, etc.). The multimodal resources activate different judges’ footings, including ANIMATOR, ANIMATOR + AUTHOR and ANIMATOR + AUTHOR + PRINCIPAL, and identify the judge’s footing shifts in the courtroom. The results also demonstrate that the judge’s footing shifts perform the functions of trial organizing, information confirming, fact investigating, spokesperson of the collegial panel, law educating and so on in criminal trials.