This study aims to develop a valid and reliable video analysis instrument to determine teachers’ character in the opening lesson in terms of its utterances and produce video analysis on micro-teaching that will be used in training related to artificial intelligence (AI). The type of research used is research and development. The subject of this research is the micro-teaching video of class E students’ batch 2019, Mathematics Education Study program at Sanata Dharma University. The product of this research is a video analysis instrument to find out the character of the prospective teachers such as confidence, enthusiasm, and happiness in terms of voice utterances. This study uses the ADDIE research model. The results show that the video analysis instrument has a validity value of 4.80 which means it’s very valid, video analysis instrument to find out the enthusiastic character has a value of validity 4.70 which means it’s very valid, video analysis instrument to find out the character of happiness has validity value of 4.60 which means it’s very valid, the practicality of the developed video analysis instrument meets the required criteria with a practicality value of 4.31 and overall, the video analysis instrument developed belongs to an effective category.