Hydrogels are highly responsive polymer networks whose swelling behavior and solvent absorption can be modulated by external stimuli. This study advances the establishment of comprehensive processing–structure–property–performance relationships for hydrogels, with a focus on deepening the structure–property relationship. We propose a data-driven approach, to derive the mathematical expression of the Flory–Huggins interaction parameter (Formula presented.) for temperature-responsive hydrogels. Starting from a presupposed expression of the interaction parameter from literature or a preliminary swelling curve observation, our method first estimates the structural parameters based on the experimental swelling data, using Bayesian optimization. These parameters are then incorporated into Flory–Rehner theory to extract the datasets of the interaction parameter depending on temperature, enabling symbolic regression to generate candidate expressions for the interaction parameter. Through multi-objective optimization, leveraging experimental swelling curves from literature, we determine sets of constants for the proposed expressions, and predict the swelling behavior. The candidate expressions are evaluated by comparing predicted and experimental swelling curves. Our approach successfully generates a superior expression for (Formula presented.), with improved agreement with the experimental data establishing the property-structure relationships in temperature-responsive hydrogels.