Accurate separation of baseflow from streamflow is of utmost importance for understanding catchment hydrological processes and supporting effective water resource management. The Smooth Minima Method is a common baseflow separation technique with a segment length parameter (<em>N</em>) representing the catchment average flow event duration. <em>N</em> is usually predicted by a power function with catchment area or default to 5 days. Yet these estimations are insufficient given the multivariate nature of <em>N</em> with other catchment attributes. In this study, we employ symbolic regression (SR) to search for possible formulation of <em>N</em> with a range of catchment attributes based on 855 catchments across the Contiguous United States. We ultimately identify three mathematical expressions of increasing complexity, achieving <em>R<sup>2</sup></em> values of 0.49, 0.50, and 0.54, compared to 0.23 and <span>−</span>0.84 for the power function and constant values. The three expressions reveal that increases exponentially with catchment area (<em>A</em>) and catchment-averaged soil saturated hydraulic conductivity (<em>K<sub>sat</sub></em>) with decreasing rates, while it increases linearly with snow day fraction (<em>f<sub>SWE</sub></em>). The effects of <em>K<sub>sat</sub></em> and <em>f<sub>SWE</sub></em> on <em>N</em> are particularly pronounced for larger values (<em>K<sub>sat</sub></em> > 25 mm/h and <em>f<sub>SWE </sub></em>> 0.4) and smaller area (<em>A</em> < 100 km<sup>2</sup>). The different calculations of <em>N</em> are also evaluated in baseflow separation, revealing higher medians of Kling-Gupta Efficiency of at least 0.84, outperforming the literature-suggested formulas for a maximum increment of 0.22. This study highlights the potential of SR for uncovering physically meaningful formulas in optimal baseflow separation.