This paper investigates the impacts of background risk on investors’ portfolio choice in a mean-variance framework and analyzes the properties of the selected portfolio and investors’ hedging behaviour. Our model implies that the optimal portfolio with background risk can be separated into two independent components: the traditional mean-variance optimal portfolio and the self-financing portfolio constructed to hedge against background risk. Our results show that both the composition and risk of the optimal portfolio are greatly affected by a number of background risk factors, including the quantity and risk of the assets that are exposed to background risk, as well as the correlation between background assets and those in the portfolio.