Diabetes or diabetes mellitus (DM) is a chronic clinical condition marked by the body's inability to regulate blood sugar (glucose) levels effectively. This condition often leads to various cardiovascular complications and can have discernible effects on electrocardiogram (ECG) signals. Diagnosis of DM typically relies on parameters such as fasting blood sugar levels, oral glucose tolerance tests, or hemoglobin A1C levels. ECG, on the other hand, is primarily employed to assess cardiac function and detect potential cardiac abnormalities that may be associated with diabetes-related complications. In this study, combination of the Short-Time Fourier Transform (STFT) and Gray-Level Co-occurrence Matrix (GLCM) techniques is explored for distinguishing between normal and diabetic ECG signals. STFT involves the transformation of 1D ECG data into a time–frequency (T–F) matrix using STFT. From each T–F matrix, four distinct GLCM features are extracted. To determine the most discriminative GLCM features for distinguishing between normal and diabetic ECG signals, we employ the Kruskal–Wallis (K–W) algorithm. We validate the effectiveness of our proposed method using an ECG database that we have compiled. The experimental findings from our research illustrate the substantial impact of this approach in identifying diabetes-related alterations in ECG signals.