This work introduces a novel quantum associative classification model that employs amplitude-encoded distance, which requires fewer qubits for data encoding and therefore uses a real value feature. This approach reduces the qubit requirements for data representation, since conventional quantum associative models often rely on binary representations of input data. The results were obtained by analyzing four balanced datasets (Iris, Cryotherapy, Data Bank Authentication, and Caesarian) and four imbalanced datasets (Haberman, Transfusion, Immunotherapy, and Fertility Diagnosis) showing that the efficiency of the quantum associative model is better than QkNN with the same metric in the case of a balanced dataset inclusive better than classical in the case of Caesarian. The results of the imbalanced show are inferior, with a performance difference of less than 10 % to that of the classical show for Haberman and Transfusion. Experiments to test simulation using IBM’s Qiskit Library.