This paper studies optimal education finance in a dynastic Mirrlees economy in which parents derive direct utility from their children’s human capital alongside standard dynastic discounting. Education-specific parental altruism adds a non-productive utility return to investment: it raises parental utility independently of the output it generates. We show that this second channel alters the constrained-efficient human-capital wedge: sufficiently strong altruism reverses the wedge from negative to positive, the optimal education subsidy is decreasing in altruism, and stronger altruism shifts intergenerational transfers away from financial bequests toward education. Calibrated to the U.S. economy, the model implies that optimal education support is non-monotonic in income and decreasing in bequests: low-income dynasties receive support due to borrowing constraints, while middle-income families face the weakest case for intervention. Income-contingent loans raise schooling, output, and welfare, but widen educational dispersion. Income-dependent subsidies reduce educational inequality more directly, at the cost of labor-supply distortions and lower aggregate output.