A novel thermally induced graft polymerization technique was used to modify a polyvinylidene fluoride (PVDF)hollow fibre microfiltration membrane. An artificial neural network (ANN) was applied to optimize the prepared condition of the membrane. The optimized dosing of acrylic acid (AA), acrylamide (AM), N, Nmethylenebisacrylamide (NMBA) and potassium persulphate (KSP) designed by ANN was that AA was 40.63 ml/L; AM acted as 6.25 g/L; NMBA was 1.72 g/L and KSP was 1.5 g/L, respectively. The thermal stability of the PVDF modified hollow fibre membrane (PVDF-PAA) was investigated by thermogravimetric (TG) and differential scanning calorimetry (DSC) analysis. The polycrystallinity of the PVDF-PAA membrane was evaluated by X-ray diffraction (XRD) analysis. The complex formation of the modified membrane was ascertained by Fourier transform infrared spectroscopy (FTIR). The morphology of the PVDF-PAA membrane was studied by environmental scanning electron microscopy (ESEM). The surface compositions of the membrane were analyzed by X-ray photoelectron spectroscopy (XPS). The adsorption capacity of Cu2+ ion on the PVDF-PAA hollow fibre membrane was also investigated.