There is no consensus on the impact of population aging on education investment. To explore this question, we first build an overlapping generations (OLG) model to theoretically analyze the effect of population aging on human capital investment in China, and then test our theory by conducting an empirical study based on micro household data. We find the following. (1) Theoretically, the OLG model shows that population aging has a crowding-out effect on education investment. (2) Empirically, the results show that the share of education and training expenditures decreases by 5.27 percentage points as the ratio of old people in the household increases by 100 percentage points, which confirms the crowding-out effect of population aging on human capital investment. (3) The crowding-out effect is far more intense on urban households than on rural households since health care expenditures will be greater in urban areas as population aging increases. (4) A quantile regression indicates that the negative effect of population aging on the share of educational expenditure is concentrated in households with higher shares of education expenditures. We confirm the robustness of our results using regional fixed effect and instrumental variable (IV) regressions.