Abstract: Morphine is not only a clinical analgesic, but also a recognized drug. Therefore, in order to achieve rapid, simple, highly sensitive and highly selective detection of morphine in urine, carboxylated multi-walled carbon nanotubes were directly applied to the surface of polished and characterized glassy carbon electrode, and the carboxylated multi-walled carbon nanotubes modified glassy carbon electrode (MWNTS-COOH/GCE) was prepared. Differential pulse adsorption stripping voltammetry (ASDPV) was used to detect morphine (MOP). The results show that after optimizing the preparation and detection conditions of the sensor, including the amount of MWNTS-COOH, enrichment time, enrichment potential, pH of buffer solution, etc., the regression equation of MWNTS-COOH/GCE working curve for morphine detection was lg(I/μA)=0.687lgc/(μmol·L⁻¹)−0.068 (r²=0.9920), and the linear range was 1.0×10⁻⁷～1.0×10⁻⁴ mol/L. The detection limit was 8.0×10⁻⁸ mol/L. After removing uric acid in the urine by uric acid enzyme, the sensor is used to detect the positive urine. Results were compared with HPLC’s, check out the absolute value of relative error within 4.24%, and through the t test found in the confidence level of 95%, there is no significant difference of the two methods.