The low photon energy and deep penetrating ability of near-infrared (NIR) light make it an ideal light source for a photoelectrochemical (PEC) immunosensing system. Absorption wavelengths of the metal-organic frameworks (MOFs) can be regulated by adjusting the metal ions and the conjugation degree of the ligands. Herein, an ionic liquid with a large conjugated structure was synthesized and was used as a ligand to coordinate with Nd ions to prepare Nd-MOF nanorods with a band gap of 1.26 eV. The Nd-MOF rods show a good photoabsorption property from 200 to 980 nm. A PEC platform was constructed by using Nd-MOF nanorods as the photoelectroactive element. A detachable double-stranded DNA labeled with alkaline phosphatase (ALP), which is specific to VEGF165, was immobilized onto the PEC sensing interface. After blocking unspecific active sites with bovine albumin, an NIR PEC aptasensing system was developed for VEGF165 detection. After being incubated in a mixture of VEGF165, l-ascorbic acid 2-phosphate (magnesium salt hydrate) (AAP), and chloroauric acid, the aptamers for VEGF165 were detached from the PEC aptasensing interface, thus resulting in the decrease of the charge-transfer resistance and the increase of the photocurrent response. The shedding of the aptamers also makes the ALP approach the electrode surface, thus catalyzing the reduction of AAP to produce ascorbic acid (AA). Subsequently, AA reduces in situ chloroauric acid to produce AuNPs on the Nd-MOF-based sensing interface. With the excellent conductivity and localized surface plasmon resonance effect, the AuNPs can accelerate the separation of electron-hole pairs generated from Nd-MOF nanorods, thus promoting the photoelectric conversion efficiency and achieving signal amplification. Under optimized conditions, the PEC responses were linearly related to the VEGF165 concentrations in the range of 0.01-100 ng mL-1 and exhibit a low detection limit of 3.51 pg mL-1 (S/N = 3). VEGF165 in human serum samples was detected by the NIR PEC aptasensor. Their concentrations were found to be well consistent with that obtained from ELISA. Furthermore, the PEC aptasensor demonstrated recoveries from 96.07 to 103.8%. The relative standard deviations were within 5%, indicating good accuracy and precision. The results further verify its practicability for clinical diagnosis.