Abstract <p>Automating and accelerating waveform design is essential for the effective adoption of piezoelectric inkjet printing (IJP). Conventionally, this goal is fundamentally restricted by a heuristic design approach that depends on labor-intensive, iterative manual tuning of voltage and timing parameters. This empirical process lacks a systematic linkage between printing parameters and the resultant waveform configuration, limiting reproducibility and scalability. To overcome these limitations, this study introduces an automatic predictive framework capable of estimating complete trapezoidal waveforms corresponding to target droplet characteristics for the given ink, thereby eliminating the need for exhaustive experimental tuning. This framework formulates the estimation problem as a multivariate regression task, enabling the simultaneous prediction of all key waveform parameters. To identify the most effective method and ensure statistical robustness, six distinct Machine Learning (ML) models were trained and evaluated. The presented results for each model are averaged over ten independent runs to guarantee stable performance. The predictive capability of the best-performing model is then rigorously validated using an entirely unseen experimental dataset. The final results exhibit strong agreement between predicted and experimental values, highlighting the framework’s potential to accelerate waveform optimization and enhance the automation of waveform design in piezoelectric IJP. <a href="https://github.com/ahsnuet/wave%20form-prediction">https://github.com/ahsnuet/wave form-prediction</a></p> Graphical abstract <p></p>