Strong, reversible, and self-cleaning adhesion in the toe pads of geckos allow the lizards to climb on a variety of vertical and inverted surfaces, regardless of the surface conditions, whether hydrophobic or hydrophilic, smooth or tough, wet or dry, clean or dirty. Development of synthetic gecko-inspired surfaces has drawn a great attention over the past two decades. Despite many external-stimuli responsive mechanisms (i.e., thermal, electrical, magnetic) have been successfully demonstrated, smart adhesives controlled by light signals still substantially lag behind. Here, in this report, we integrate tetramethylpiperidinyloxyl (TEMPO)-doped polydopamine (PDA), namely, TDPDA, with PDMS micropillars using a template-assisted casting method, to achieve both improved adhesion and self-cleaning performances. To the best of our knowledge, this is the first report on PDA being used as a doping nanoparticle in bioinspired adhesive surfaces to achieve highly efficient self-cleaning controllable by light signals. Notably, the adhesion of the 5% TDPDA-PDMS sample is ∼688.75% higher than that of the pure PDMS at the individual pillar level, which helps to explain the highly efficient self-cleaning mechanism. The sample surfaces (named TDPDA-PDMS) can efficiently absorb 808 nm wavelength of light and heat up from 25 °C to 80.9 °C in 3 min with NIR irradiation. The temperature rise causes significant reduction of adhesion, which results in outstanding self-cleaning rate of up to 55.8% within five steps. The exploration of the photoenabled switching mechanism with outstanding sensitivity may bring the biomimetic smart surfaces into a new dimension, rendering varied applications, e.g., in miniaturized climbing robot, artificial intelligence programmable manipulation/assembly/filtration, active self-cleaning solar panels, including high output sensors and devices in many engineering and biomedical frontiers.