Complex surfaces in advanced engineering materials (AEMs) can be generated using abrasive waterjet (AWJ) technology due to the high material removal rate, high surface quality and hence can be utilized in the healthcare, aerospace, and automotive industries. Although AWJ technology can machine AEMs without burr, long chips, and thermal deformation, unlike conventional machining, the process is stochastic aggressive, and producing high-quality finish is challenging. Furthermore, the surface has irregular depth and waviness due to process volatility and material ductility. Hence, a characterization approach is needed to measure the quality of milled parts. Although the confocal laser scanning microscope and optical profilometer are used to measure texture characteristics, these have drawbacks, such as (i) the difficulty of incorporating into the production workflow and providing real-time feedback and (ii) the high cost and slow operational speed. Therefore, the present work proposes an in-situ approach for predicting the surface waviness (Wa) of the AWJ milled components. The proposed approach leverages machine learning (ML) techniques to analyze the image data during milling and extract valuable insights. Furthermore, the principal component analysis (PCA) is incorporated to determine the most influential characteristics in the data, allowing for a better understanding of the underlying variables influencing surface quality and reducing the multi-collinearity effect. Following this, the ML model is used to train the PCA features to predict surface waviness. Among all the prediction models considered in the study, Huber’s regressor model provided the most accurate prediction, with an R2 value of 0.92 for surface waviness.