There has been much effort on predicting users' location (country or region) using social media sites, like Twitter. However, exciting work has not addressed the prediction of users' place of visit at finer granularity like restaurant or park. To address this problem, in this paper, we present the methodology of predicting user's place of visit using: (i) personality attributes of users obtained from her tweets using SystemU; and (ii) time and day at which prediction is to be made. Based on these features, we build a model that predicts the likelihood of a user's place of visit at given time and day. We performed extensive experiments involving, real data derived from twitter timelines of more than 3000 users. Models are trained using different machine learning algorithms and results demonstrate the effectiveness of work with accuracy more than 80% for top 5 predictions.