In this article, we review our recent studies on the ultrafast vibrational dynamics at aqueous interfaces carried out with two-dimensional (2D) heterodyne-detected vibrational sum frequency generation (HD-VSFG) spectroscopy. Compared to the wealth of knowledge about bulk water, molecular-level understanding of interfacial water is still poor due to the technical difficulty in selectively observing molecules at the interfaces. HD-VSFG spectroscopy is based on the second-order optical process and thus intrinsically interface-selective. 2D HD-VSFG spectroscopy is its extension to the time-resolved measurement, and it is an interfacial analog of 2D IR spectroscopy which has been extensively utilized for bulk studies. This novel interface-selective ultrafast spectroscopy has enabled us to investigate ultrafast vibrational dynamics at aqueous interfaces at the high level equivalent to the bulk studies. We describe the principle and instrumentation of 2D HD-VSFG spectroscopy as well as several selected examples of 2D HD-VSFG studies that provided new insights into aqueous interfaces. At the air/neat water interface, 2D HD-VSFG indicated high similarity of hydrogen-bonded OH of interfacial water to that of bulk water while unique non-hydrogen bonded OH is present at the interface. At the charged surfactant/water interfaces, 2D HD-VSFG enabled us to clearly observe ultrafast spectral diffusion in the OH stretch band and demonstrated the importance of isotopic dilution for unambiguous observation of vibrational dynamics. At model membrane lipid/water interfaces, it was found that the hydrogen-bonded dynamics is greatly affected by the interaction between the interfacial water and the head group of the lipids and that the effects of coexisting head groups cannot simply be summed up but they are highly cooperative.