<p>In this paper, we study an international portfolio selection problem, which allocates wealth in different security markets. We built an international portfolio selection model with a chance constraint to guarantee the portfolio performance over a benchmark in a large probability. Stock returns are modeled by a multi-factor structure with Gaussian residuals, while factors, exchange rates and the benchmark follow heavy-tailed marginals coupled by copulas to capture nonlinear dependence. We develop a partial sampling approximation method combined with a sequential convex approximation method to solve the chance constrained international portfolio selection problem. Extensive experiments on developed and emerging markets show the reasonability and superior out-of-sample performance of the proposed international portfolio selection model.</p>