Sensitive and selective detection of ultralow concentrations of specific biomolecules is important in early clinical diagnoses and biomedical applications. Many types of aptasensors have been developed for the detection of various biomolecules, but usually suffer from false positive signals and high background signals. In this work, we have developed an amplified fluorescence aptasensor platform for ultrasensitive biomolecule detection based on enzyme-assisted target-recycling signal amplification and graphene oxide. By using a split molecular aptamer beacon and a nicking enzyme, the typical problem of false positive signals can be effectively resolved. Only in the presence of a target biomolecule, the sensor system is able to generate a positive signal, which significantly improves the selectivity of the aptasensor. Moreover, using graphene oxide as a super-quencher can effectively reduce the high background signal of a sensing platform. We select vascular endothelial growth factor (VEGF) and adenosine triphosphate (ATP) as model analytes in the current proof-of-concept experiments. It is shown that under optimized conditions, our strategy exhibits high sensitivity and selectivity for the quantification of VEGF and ATP with a low detection limit (1 pM and 4 nM, respectively). In addition, this biosensor has been successfully utilized in the analysis of real biological samples.