Cardiac arrhythmias are time-varying, non-period, irregular, or complex fluctuations in time-domain waveforms, revealing multiple morphologies, including supraventricular, junctional, and ventricular arrhythmias, and conduction abnormalities. Given that visual examinations are time-consuming, this study proposes a screening method for identifying the classes of multiple cardiac arrhythmias using the discrete fractional-order integration (DFOI) and the perceptual color representation-based intelligent classifier. The DFOI operation with the specific fractional order is employed to extract the QRS features with the finite computations and finite power series in a specific timing duration. DFOI can deal with the non-period and irregular time-varying signals. The proposed intelligent classifier consists of Bayesian network and color relation analysis (CRA). This classifier is employed to identify the ectopic classes with perceptual color representation. Using the arrhythmia database of Massachusetts Institute of Technology-Beth Israel Hospital database, the proposed intelligent classifier demonstrated high efficiency in recognizing electrocardiography signals.