During the regeneration of the oxidized dye in dye-sensitized solar cells, the redox couple of I-/I-3(-) reduces the photo-oxidized dye. The simplest mechanism would be a direct charge-transfer mechanism from I- to D+ [D+ + I- -> D-0 + 1], called the single iodide process (SIP). However, this is an unfavorable equilibrium because the redox potential of I*/I- is 1.224 V vs SHE, which is 0.13 V higher than that of the dye. This led to the postulation of the two iodide process (TIP) [(D+...I-) + I- -> (D...I-2(-)) -> D-0 + I-2(-)] for a sufficiently high reducing power, but TIP is not consistent with either the recent experimental data suggesting the first-order kinetics or recent time-resolved spectroscopic measurements. To resolve this conundrum, we used quantum mechanics including Poisson-Boltzmann solvation to examine the electron-transfer process between I- and D+ for the Ru(dcb)(2)NCS2 or N3 dye. We find that I- is attracted to the oxidized dye, positioning I- next to the NCS. At this equilibrium position, the I- electron is already 40% transferred to the NCS, showing that the redox potential of I- is well matched with the dye. This matching of the redox potential occurs because I- is partially desolvated as it positions itself for the inner-sphere electron transfer (ISET). The previous analyses all assumed an outer-sphere electron-transfer process. Thus our ISET-SIP model is consistent with the known redox potentials and with recent experimental reports. With the ISET-SIP mechanism, one can start to consider how to enhance the dye regeneration kinetics by redesigning ligands to maximize the interaction with iodide.