The S-O-R theory by Hovland, et.al (1953) states that organisms produce certain behaviors if there are special stimulus conditions, so that one can expect and estimate the compatibility between messages and communicant reactions (Effendy, 2005: 254). Twitter exposure can affect various aspects of the lives of its users, one of which is in the aspect of language attitudes. Language attitude is defined as the subjective perception of language influenced by behaviors. This research aims to analyze the effect of twitter exposure on language attitude of college students in Brawijaya University. This research is quantitative research. The data were collected using a questionnaire, from the results of the questionnaire were analyzed to determine the effect of each variable calculated using regression analysis. To convert data from questionnaires into numerical data that can be analysed, a Likert scale is used. This study shows the influence of twitter media exposure (X) on language attitude (Y) is significant. In addition, t value is -10.771 which means that the influence is inversely proportional, the more twitter media exposure increase, language attitude will decrease. This means that H0 is rejected, so it can be concluded that language attitude (Y) can be significantly influenced.