Theoretical Bit Error Rate (BER) and channel capacity analysis are always of great interest to the designers of wireless communication systems. At the center of such analyses people are often encountered with a high-dimensional multiple integrals with quite complex integrands. Conventional Gaussian quadrature is inefficient in handling problems like this, as it tends to entail tremendous computational overhead, and the principal order of its error term increase rapidly with the dimension of the integral. In this paper, we propose a new approach to calculate complex multi-fold integrals based on the number theory. In contrast to Gaussian quadrature, the proposed approach requires less computational effort, and the principal order of its error term is independent of the dimension. The effectiveness of the number theory based approach is examined in BER and capacity analyses for practical systems. In particular, the results generated by numerical computation turn out in good match with that of Monte-Carlo simulations.