Flexibly reconfigurable roll forming (FRRF) is a sheet-forming technology that can be used to produce multi-curvature surfaces by controlling the longitudinal strain distribution. In FRRF, the shape of the formed surface is determined by the curvature of the reconfigurable rollers and the gaps between the rollers. Because FRRF technology is still under development, a simulation model of the physical forming process is conveniently used to investigate the effects of the input parameters. To facilitate the investigation in the present study, the response surface methodology is used to develop a model for predicting the curvature produced in a longitudinal blank. The input parameters are the sheet compression ratio, the curvature radius in the transverse direction, and the initial blank width. Samples are generated using a three-level three-factor full-factorial design, and each convex and saddle curvature is represented by a quadratic regression model with two-factor interactions. The fitted polynomial equations are verified numerically by the R-squared values and root mean square errors and graphically by residual plots. To assess the reliability of the sample data, experiments are performed using pre-FRRF equipment. The proposed analytical procedure is confirmed to be reasonable, and a statistical formula for estimating the longitudinal curvature produced by the FRRF process is established.