This study uses patent analysis to explore the corporate performance of the American pharmaceutical industry by use of artificial neural network (ANN). Besides, this paper explores the nonlinear influences of patent-counts, self-citations, relative patent position upon corporation performance of the American pharmaceutical companies. The results show that relative patent position has a nonlinearly and monotonically positive influence upon the corporation performance. However, patent counts and self-citations have inverse U-shaped influences upon corporation performance. Therefore, if the American pharmaceutical companies want to enhance their performance, they should enhance their relative patent position in their most important technological fields and adjust their patent counts and self-citations to the optimal points.