Recently we described an association between psoriasis and KIR2DS1, a gene for a stimulatory natural killer cell receptor, in a Polish population. The association was independently reported among Japanese and confirmed in a U.S. population. Prompted by these findings, we reanalyzed data by a multivariate approach in search of possible effects of KIR genes other than KIR2DS1 (non-KIR2DS1). The methodology was based on a stratified analysis and multiple logistic regression. We found that the non-KIR2DS1 genes had joint effects comparable to or stronger than the effects of KIR2DS1 in both the fraction of explained variance (0.174 vs 0.204, respectively, for KIR2DS1 and non-KIR2DS1) and the statistical significance (p = 0.000008 vs p = 0.000001, respectively). When individual genes were considered, a decrease in KIR2DS5 among patients vs controls (OR = 0.2, pcor = 0.0005) and a decrease in KIR2DS3 restricted to KIR2DS1-positive individuals (OR = 0.2, pcor = 0.005) were evident. We also performed a multivariate analysis of the HLA-Cw genotypes but failed to demonstrate any effects in addition to the known association with HLA-Cw*06. We conclude that the effect of the KIR genes on psoriasis susceptibility is complex, extending beyond the association with KIR2DS1 and involving protective effects and interactions.