Recently, ternary blend polymer solar cells have attracted much attention because they can harvest the sun light over a wide wavelength range from visible to near-IR region.  As a result, photovoltaic performances of polymer solar cells have been improved significantly.  In this talk, I will talk about three types of ternary blend polymer solar cells and discuss how to improve the photovoltaic performance.  First, I will talk about ternary blend polymer solar cells based on a wide-bandgap conjugated polymer, a fullerene acceptor, and a near-IR dye.  In this case, dye molecules should be located at a donor/acceptor interface in ternary blend solar cells.  Such a dye location can be controlled by careful tuning of surface energy of three materials.  We found that hetero structured dye molecules with different ligands are spontaneously located at the interface and hence can improve photocurrent generation most effectively.  Second, I will talk about ternary blend polymer solar cells based on a wide-bandgap conjugated polymer, a low-bandgap polymer, and a fullerene acceptor.  Because of complimentary absorption bands of the two donor polymers, the photocurrent can be improved in comparison with that of binary blend counterpart solar cells.  Furthermore, a fill factor is also increased even for the ternary blend solar cells, suggesting that charge transport is improved in ternary blends rather than in binary blends.  Third, I will talk about ternary blend polymer solar cells based on a wide-bandgap conjugated polymer, a low-bandgap conjugated polymer, and a non-fullerene acceptor. Similarly, the photocurrent and fill factor are improved in ternary blend solar cells.  Interestingly, the photocurrent conversion efficiency can be as high as ~80% even at a near-IR wavelength region. As a result, more than 11% efficiency was obtained.  This finding suggest that non-fullerene acceptors are suitable as near-IR harvesting materials.