All-polymer solar cells (all-PSCs), consisting of conjugated polymers as both electron donor (P) and acceptor (P), have recently attracted great attention. Remarkable progress has been achieved during the past few years, with power conversion efficiencies (PCEs) now approaching 8%. In this Account, we first discuss the major advantages of all-PSCs over fullerene-polymer solar cells (fullerene-PSCs): (i) high light absorption and chemical tunability of P, which affords simultaneous enhancement of both the short-circuit current density (J) and the open-circuit voltage (V), and (ii) superior long-term stability (in particular, thermal and mechanical stability) of all-PSCs due to entangled long P chains. In the second part of this Account, we discuss the device operation mechanism of all-PSCs and recognize the major challenges that need to be addressed in optimizing the performance of all-PSCs. The major difference between all-PSCs and fullerene-PSCs originates from the molecular structures and interactions, i.e., the electron transport ability in all-PSCs is significantly affected by the packing geometry of two-dimensional P chains relative to the electrodes (e.g., face-on or edge-on orientation), whereas spherically shaped fullerene acceptors can facilitate isotropic electron transport properties in fullerene-PSCs. Moreover, the crystalline packing structures of P and P at the P-P interface greatly affect their free charge carrier generation efficiencies. The design of P polymers (e.g., main backbone, side chain, and molecular weight) should therefore take account of optimizing three major aspects in all-PSCs: (1) the electron transport ability of P, (2) the molecular packing structure and orientation of P, and (3) the blend morphology. First, control of the backbone and side-chain structures, as well as the molecular weight, is critical for generating strong intermolecular assembly of P and its network, thus enabling high electron transport ability of P comparable to that of fullerenes. Second, the molecular orientation of anisotropically structured P should be favorably controlled in order to achieve efficient charge transport as well as charge transfer at the P-P interface. For instance, face-to-face stacking between P and P at the interface is desired for efficient free charge carrier generation because misoriented chains often cause an additional energy barrier for overcoming the binding energy of the charge transfer state. Third, large-scale phase separation often occurs in all-PSCs because of the significantly reduced entropic contribution by two macromolecular chains of P and P that energetically disfavors mixing. In this Account, we review the recent progress toward overcoming each recognized challenge and intend to provide guidelines for the future design of P. We believe that by optimization of the parameters discussed above the PCE values of all-PSCs will surpass the 10% level in the near future and that all-PSCs are promising candidates for the successful realization of flexible and portable power generators.