Strong evidence of linkage between chromosomal region 6q16-q22 and bipolar affective disorder (BPAD) has previously been reported. We conducted a systematic association mapping of the 6q-linkage interval using 617 SNP markers in a BPAD case-control sample of German descent (cases = 330, controls = 325). In this screening step, 46 SNPs showed nominally significant BPAD-association (P-values between 0.0007 and 0.0484). Although none of the 46 SNPs survived correction for multiple testing, they were genotyped in a second and ethnically matched BPAD sample (cases = 328, controls = 397). At the melanin-concentrating-hormone-receptor-2 (MCHR2) gene, we found nominal association in both the initial and second BPAD samples (combined P = 0.008). This finding was followed up by the genotyping of 17 additional MCHR2-SNPs in the combined sample in order to define our findings more precisely. We found that the MCHR2-locus can be divided into three different haplotype-blocks, and observed that the MCHR2-association was most pronounced in BPAD male patients with psychotic symptoms. In two neighboring blocks, putative risk-haplotypes were found to be 7% more frequent in patients (block II: 23.3% vs. 16.2%, P = 0.005, block III: 39.2% vs. 32.0%, P = 0.024), whereas the putative protective haplotypes were found to be 5-8% less frequent in patients (block II: 11.6% vs. 16.4%, P = 0.041, block III: 30.0% vs. 38.8%, P = 0.007). The corresponding odds ratios (single-marker analysis) ranged between 1.25 and 1.46. Our findings may indicate that MCHR2 is a putative risk factor for BPAD. These findings should be interpreted with caution and replicated in independent BPAD samples.