Abstract: Electrostatic discharge (ESD) is a key factor in the safe operation of electronic devices, with its waveform characteristics significantly influenced by various parameters, such as approach speed. To address this, the present study innovatively developed an electrode motion speed effect tester to collect ESD waveform data. The raw data were preprocessed using equal-width binning and various normalization methods, and time-domain, frequency-domain, and morphological features were extracted. Based on these features, multiple regression prediction models were constructed, including traditional machine learning methods (support vector machine (SVM), random forest, decision tree, and k-nearest neighbors) as well as deep learning methods (long short-term memory neural network (LSTM), one-dimensional convolutional neural network, Transformer, and Transformer+LSTM) for the multi-parameter prediction of ESD parameters (such as peak value and rising rate). Experimental results indicate that, on the dataset based on Min-Max normalization, the SVM model achieved the best performance in predicting ESD parameters, with determination coefficients of 0.9913 and 0.9829 respectively, and significantly lower mean squared error and mean absolute error compared to other methods. This demonstrates that SVM exhibits high robustness and generalization ability when handling small sample sizes and high-dimensional data, while deep learning models show potential advantages in automatic feature extraction. The innovative aspect of this study lies in its systematic comparison of multiple traditional and deep learning models, providing a novel perspective for ESD parameter prediction and offering theoretical and practical guidance for the optimization and application of electrostatic protection technology.