Family and twin studies of Borderline Personality Disorder (BPD) have found familial aggregation and genetic propensity for BPD, but estimates vary widely. Large-scale family studies of clinically diagnosed BPD are lacking. Therefore, we performed a total-population study estimating the familial aggregation and heritability of clinically diagnosed BPD. We followed 1,851,755 individuals born 1973-1993 in linked Swedish national registries. BPD-diagnosis was ascertained between 1997 and 2013, 11,665 received a BPD-diagnosis. We identified relatives and estimated sex and birth year adjusted hazard ratios, i.e., the rate of BPD-diagnoses in relatives to individuals with BPD-diagnosis compared to individuals with unaffected relatives, and used structural equation modeling to estimate heritability. The familial association decreased along with genetic relatedness. The hazard ratio was 11.5 (95% confidence interval (CI) = 1.6-83.8) for monozygotic twins; 7.4 (95% CI = 1.0-55.3) for dizygotic twins; 4.7 (95% CI = 3.9-5.6) for full siblings; 2.1 (95% CI = 1.5-3.0) for maternal half-siblings; 1.3 (95% CI = 0.9-2.1) for paternal half-siblings; 1.7 (95% CI = 1.4-2.0) for cousins whose parents were full siblings; 1.1 (95% CI = 0.7-1.8) for cousins whose parents were maternal half-siblings; and 1.9 (95% CI = 1.2-2.9) for cousins whose parents were paternal half-siblings. Heritability was estimated at 46% (95% CI = 39-53), and the remaining variance was explained by individually unique environmental factors. Our findings pave the way for further research into specific genetic variants, unique environmental factors implicated, and their interplay in risk for BPD.