<p>Analyzing single-molecule binding kinetics offers an effective way to reduce the disturbance from nonspecific bindings in biosensors. Here we present a dual-parameter lifetime distribution modeling approach to detect specific binding signals in single-molecule sensors accurately. A proof of concept was demonstrated in the dynamic single-molecule sensing of microRNA using gold-nanoparticle labeled sandwiched assay with low-affinity probes. In this assay, a single molecule binding process was recorded with a large field-of-view plasmonic scattering microscope, and the lifetime distribution was quantified. A model involving two different exponential decaying constants was used to fit the lifetime distribution, providing accurate information about the number of nonspecific binding and specific binding events as well as their dissociation rates. We show both in simulations and experiments that by establishing the calibration curve with the number of specific binding events against different analyte concentrations, an ultra-low limit of detection at the femtomolar level could be achieved. The high sensitivity makes this approach a potential solution to detect low abundance biomarkers for disease diagnosis.</p>