Objective: We aimed to examine whether variability in high-density lipoprotein cholesterol (HDL-c) over time was associated with cognitive function. Method: We conducted a post hoc analysis of the PROspective Study of Pravastatin in the Elderly at Risk (PROSPER) trial. Our sample included 4,428 participants with at least two repeated HDL-c measures between Months 3 and 24 postbaseline and with cognitive assessments at Month 30. HDL-c variability was defined as the intraindividual standard deviation over each person’s repeated measurements. Results: Higher HDL-c variability was associated with worse performance on the Letter-Digit Coding Test (β [95% confidence interval] [CI] = −4.39 [−7.36, −1.43], p = .004), immediate recall on the 15-Picture Learning Test (β [95% CI] = −0.98 [−1.86, −0.11], p = .027), and delayed recall on the 15-Picture Learning Test (β [95% CI] = −1.90 [−3.14, −0.67], p = .002). The associations did not vary by treatment group. Discussion: Our findings suggest that variability in HDL-c may be associated with poor cognitive function among older adults.