Polymer-based solar cells have made rapid progress in the last decade and are currently attracting a great deal of attention as a next-generation solar cell. Very recently, they have shown a power conversion efficiency (PCE) of more than 10%, which is comparable to that reported for amorphous silicon solar cells. However, it is still required to improve the photovoltaic performance furthermore for practical applications. In this talk, I will demonstrate ternary blend polymer solar cells, which are a new approach to improving the photocurrent generation. We have recently developed ternary blend polymer solar cells based on a wide-bandgap polymer, poly(3-hexylthiophene) (P3HT), a fullerene derivative (PCBM), and a near-IR dye molecule such as a silicon phthalocyanine derivative (SiPc). Such near-IR dye addition can easily expand the light-harvesting wavelength range up to the near-IR region, and hence can boost the photocurrent furthermore. I will demonstrate how molecular design of near-IR dye molecules can control the location in ternary blend solar cells to improve the photocurrent generation effectively. We also have fabricated ternary blend polymer solar cells based on a wide-bandgap polymer, a low-bandgap polymer, and PCBM. This two-donor polymer blend can also expand the light-harvesting wavelength and hence can boost the photocurrent effectively. Efficient exciton harvesting and charge collection can be designed by combining a wide-bandgap crystalline polymer and a low-bandgap amorphous polymer.