Many studies in Western societies show a pattern of discriminative grandparental investment as follows: maternal grandmothers (MGMs) > maternal grandfathers (MGFs) > paternal grandmothers (PGMs) > paternal grandfathers (PGFs). This pattern is in line with the expectation from evolutionary reasoning. Yet whether or not this pattern applies in China is in question. The present study was based on a questionnaire survey at a university in Central China (N = 1,195). Results show that (1) when grandparent-grandchild residential distance during grandchildren's childhood is controlled, in the case of grandsons and granddaughters as a whole and granddaughters only, both grandparental caregiving and grandchildren's emotional closeness to grandparents display a rank order of MGM > MGF > PGM > PGF, but in the case of grandsons only, this order is not statistically significant. (2) There are stable relationships between grandparental caregiving/grandchildren's emotional closeness and residential distance/similarity in appearance. (3) The effects of residential distance on either PGFs' or PGMs' caregiving exceed those on either MGFs' or MGMs'. (4) The PGF and PGM prefer grandsons to granddaughters in their caregiving, whereas the MGF and MGM do not have a sex preference, and (5) the fact that the PGF and PGM invest more in grandsons than in granddaughters does not depend on grandsons' duration of living in a rural area. Our results suggest that (1) in general, the Chinese display a pattern of differential grandparental investment predicted by an evolutionary perspective, (2) the evolutionary perspective that combines the two factors of paternal uncertainty and sex-specific reproductive strategies is applicable to grandparental investment in China, and (3) the traditional son-preference culture also plays some role in affecting grandparental investment in China, though the roles of culture and urban-rural cultural difference should not be exaggerated.