An urgency to do a reverse engineering of predicting required reagents from a given polymer gelation time is paramount to support an efficient and economical water conformance control in petroleum field. The gelation period should consider the duration of solution pumping, desired well shut-in, and well reopening. There are two major problems could occur in water shut-off treatment: a premature or a long-waited gelling. To address those challenges, we conduct a comprehensive data compilation, review on polymer gelation models, assessment of the state-of-the-art in responsive polymers, and gelation time predictions based on our own developed analytical models and selected machine learning algorithms. Physics-based and data-driven study for a polyacrylamide-hexamethylenetetramine-hydroquinone system reveal that higher temperatures, reagents concentrations, and pH induce quicker gelation processes. Hexamethylenetetramine is salient crosslinking agent compared to hydroquinone. Contrarily, the incremental of salinity prolong the gelation time. Predictions on polymer gelation time with Random Forest algorithm perform better among Artficial Neural Network, Convolutional Neural Network, and K-Nearest Neighbor.