With the advent of the era of big data, the world is facing the challenge of unprecedented information blowout growth, and Internet information processing technology has also ushered in rapid development. At the same time, charity culture has always been a part that the state and society attach great importance to. It is a form of social wealth redistribution and a supplement to the social security system. Based on this social background and reality, this paper analyzes and explores the psychological changes of charitable cultural behavior by social media based on big data technology. This paper introduces latent semantic analysis and constructs the theoretical framework of social media. This paper introduces the related concepts of heart rate variability and mental state, collects the factor factors and constructs the model, and studies the PLS load and cross load. This paper makes a statistical analysis of the data of charitable donations in China from 2017 to 2021 and tests the validity of the questionnaire from the psychological level. The experimental results show that the correlation between the subscales of the questionnaire and the total scale is between 0.470 and 0.820, and the correlation between the dimensions is between 0.183 and 0.590. The correlation between the dimensions and the total score of the scale is higher than that between the dimensions, which shows that there are both connections and differences between the dimensions of the two scales, which are relatively independent, and the scale has a certain structural validity.