Abstract: With the outgrowth of user-based web content, individuals can freely express their opinion in many domains. However, this would imply a huge cost to annotate training data for a large number of domains and prevent us from exploiting the information shared across various domains. As a result, cross-domain sentiment analysis is a challenging NLP task due to feature divergence and polarity divergence. However, to tackle this issue, this study presents a new model for cross-domain sentiment classification. This model is based on transferring features between source and target domains vice versa, using a Union of Conditional Probability (UCP) association measure. A Naive Bayes (NB) classifier and three feature selection methods (Information gain, Odd ratio, Chi-square) are used to evaluate the proposed model. Experimental results show that our model’s results were very promising and encourages us to further pursue this research.