In this study, a panel of 13 STR loci locate on chromosome 3, 4, and 17 (D3S2402, D3S2452, D3S1766, D3S4554, D3S2388, D3S3051, D3S3053, D4S2404, D4S2364, AC001348A, AC001348B, D17S975, and D17S1294) were assessed for pairwise kinship analysis. Map distances between these STR loci ranged from 0.07 cM to 97.03 cM. The population genetic study of Chinese Han population showed that linkage disequilibrium exists in two clusters of closely linked markers (D4S2404-D4S2364 and D17S975-D17S1294), in which the recombination fractions were 0.0026 and 0.0001, respectively. The recombination fractions derived from the Rutgers Map for the closely linked markers (genetic distance < 0.5 cM) were significant underestimates in comparison to those of direct observation of STR transmissions in families. When effect of linkage on pairwise kinship testing was evaluated by comparing likelihood ratio (LR) values taking linkage into account, overall LR values increased. But extremely low LRs were also observed. Finally, the power of the 13 STR loci to discriminate relationship among full-sibs, half-sibs, grandparent-grandchild, uncle-niece, and unrelated pairs was assessed with a category fraction. The results showed that about 72.64% of full-sib pairs and about 82.84% of unrelated pairs could be classified correctly. But the category fractions of second-degree relationships drastically reduced to 7.34-35.48%. If only pairs of grandparent-grandchild, half-sibs, and uncle-niece were distinguished, the category fraction was 0.5512, 0.1147, and 0.4362, respectively. Our study results demonstrated that linked STRs were helpful to differentiate the most frequent relationships in pairwise kinship analysis.