Fludarabine is the most frequently used agent in conditioning regimens for allogeneic hematopoietic cell transplantation (HCT). Body surface area-based dosing leads to highly variable fludarabine exposure. We studied the relation between fludarabine exposure and clinical outcomes. A retrospective, pharmacokinetic-pharmacodynamic analysis was conducted with data from patients undergoing HCT with fludarabine (160 mg/m2) as part of a myeloablative conditioning (busulfan targeted to an area under the plasma-concentration-time curve [AUC] of 90 mg*h/L) and rabbit antithymocyte globulin (6-10 mg/kg; from day -9/-12) between 2010 and 2016. Fludarabine exposure as AUC was calculated for each patient using a previously published population pharmacokinetic model and related to 2-year event-free survival (EFS) by means of (parametric) time-to-event models. Relapse, nonrelapse mortality (NRM), and graft failure were considered events. One hundred ninety-two patients were included (68 benign and 124 malignant disorders). The optimal fludarabine exposure was determined as an AUC of 20 mg*h/L. In the overexposed group, EFS was lower (hazard ratio [HR], 2.0; 95% confidence interval [CI], 1.1-3.5; P = .02), due to higher NRM (HR, 3.4; 95% CI, 1.6-6.9; P <001) associated with impaired immune reconstitution (HR, 0.43; 95% CI, 0.26-0.70; P <001). The risks of NRM and graft failure were increased in the underexposed group (HR, 3.3; 95% CI, 1.2-9.4; P = .02; HR, 4.8; 95% CI, 1.2-19; P = .02, respectively). No relationship with relapse was found. Fludarabine exposure is a strong predictor of survival after HCT, stressing the importance of optimum fludarabine dosing. Individualized dosing, based on weight and "renal function" or "therapeutic drug monitoring," to achieve optimal fludarabine exposure might improve survival.