The two-step digital-to-analog conversion (DAC) scheme has been reported to be very area-efficient for thin-film transistor-liquid crystal display (TFT-LCD) data driver ICs, but it is not as well suited as it is for polycrystalline silicon (poly-Si) TFT integrated circuits. The charge redistribution in the two-step DAC process requires an operational amplifier in principle, which is most challenging for poly-Si TFT circuits. The proposed two-step DAC for active-matrix organic light-emitting diodes (AMOLEDs) makes the operational amplifier unnecessary by appropriately exploiting the preexisting capacitors in the pixel to compensate for the characteristic variations of TFTs. Moreover, the second-step DAC occurs at the same time as threshold voltage compensation, and it does not require additional time. By thoroughly analyzing area efficiency depending on the resolution decomposition between the first- and second-step DACs, we found that 5-bit coarse plus 3-bit fine DACs are best in terms of circuit area. When we designed a layout of the proposed 8-bit DAC on the basis of the 4 µm design rules, the DAC circuit area is no more than 72×637 µm2, which, to the best of our knowledge, is the most compact to date.