Inkjet printing offers significant potential for additive manufacturing technology. However, predicting jetting behavior is challenging because the rheological properties of functional inks commonly used in the industry are overlooked in printability maps that rely on the Ohnesorge and Weber numbers. We present a machine learning-based predictive model for jetting behavior that incorporates the Deborah number, the Ohnesorge number, and the waveform parameters. Ten viscoelastic inks have been prepared and their storage modulus and loss modulus measured, showing good agreement with those obtained by the theoretical Maxwell model. With the relaxation time of the viscoelastic ink obtained by analyzing the Maxwell model equations, the Deborah number could be calculated. We collected a large data set of jetting behaviors of each ink with various waveforms using drop watching system. Three distinct machine learning models were employed to build predictive models. After comparing the prediction accuracy of the machine learning models, we found that multilayer perceptron showed outstanding prediction accuracy. The final predictive model exhibited remarkable accuracy for an unknown ink based on waveform parameters, and the correlation between jetting behavior and ink properties was reasonable. Finally, we developed a printability map characterized by the Ohnesorge and Deborah numbers through the proposed predictive model for viscoelastic fluids and the chosen industrial printhead.