Emotion has always been a complex psychological and physiological phenomenon of people. In order to solve the problem of emotional calculation and emotional measurement methods, it will make people's research and understanding of emotion more meaningful in the future. This article aims to study the research of emotional computing and emotional measurement methods based on intelligent algorithms for wireless network communication. This paper proposes a modelling algorithm of physiological signal emotion based on random forest and LDA. After extracting the general features and proprietary features of the signal, the random forest algorithm is used to calculate the importance of these features under different labels, and the effective features are selected from all the features according to the order of importance. This article also realizes the process of establishing a complete emotional material database and physiological emotional database. Through multiple steps such as signal screening, material selection, emotion annotation and physiological data collection, the problem of missing physiological emotion database is solved. Experimental data shows that compared with other low-level feature classification, the recognition rate of the fusion method in this paper is 57.37% higher than other algorithms, the recognition rate of other behaviours is increased by 4.35%.