Due to its low-input requirements, high yield capacity in marginal soils, and high carbohydrate and vitamin A content, sweetpotato is an important food security crop. For effective breeding strategies, knowing and understanding the traits that drive diversity among varieties is required. In this study, a set representing the diversity of sweetpotato varieties was characterized through multivariate ordination based on growth parameters and physiological leaf traits. The dynamic of light interception, light use efficiency (LUE) parameters, and partition of assimilates to the storage roots of eight representative varieties were simulated through a crop growth model. Leaf mass per area (LMA), N and P content in leaves, light response curve parameters and carbon discrimination were assessed in potted plants at early-growing period (58 days after planting). Precocity proxies (inversely related to LMA and thermal time at maximum storage root growth) and conversion efficiency of intercepted radiation indicators (quantum yield and LUE) were important factors in the varieties ordination based on leaf traits and growth parameter scales, respectively. Leaf traits assessment at early stages could be used as a starting point for the screening of potential lines, which once identified can be further characterized using crop growth models.