Purpose: Speech sound disorder (SSD) is one of the major speech disorders in school-age children. Given the heterogeneity in terms of subtypes within SSD, there is a need to develop techniques for a quick identification of these subtypes. Furthermore, given the paucity of studies from children with SSD from Cantonese-speaking homes and a noted prevalence of SSDs in Cantonese-speaking children, it becomes even more important to investigate the subtypes of SSDs in Cantonese-speaking children. In the current study, using a combined traditional ranking-based and novel weightage-based optimality theory (OT) approach, we conducted an inquiry in Cantonese-speaking 3- to 6-year-olds with and without SSD. Method: We compared the speech sound productions from 31 children with SSD (3 years old: n = 12; 4 years old: n = 9; 5 years old: n = 10) with 30 typically developing children (3 years old: n = 9; 4 years old: n = 10; 5 years old: n = 11). Speech samples were analyzed using a ranking-based and weightage-based OT approach. Results: Using the markedness hierarchy among affricates, fricatives, and plosives, we found that 77.4% of children in the SSD group conformed to the "delay" subtype, while 22.6% of children within the SSD group conformed to the "atypical" subtype. More specifically, for the typically developing children and the SSD-delay subgroup, stopping, de-affrication, and de-aspiration were observed, and the weight difference between faithfulness and markedness constraints increased with age. On the contrary, for the SSD-atypical subgroup, frication, affrication, and aspiration were found, and the values of weight difference between constraints decreased with age. Conclusions: The findings from the current study suggest that the weightage-based Maximum Entropy grammar approach can delineate between the SSD subgroups (i.e., SSD-delay vs. SSD-atypical) defined by the ranking-based traditional OT approach. These findings offer a starting point into the development of objective tools for clinicians for detecting the SSD subgroups to make decisions on treatment type, as we speculate different treatment approaches for SSD-delay versus SSD-atypical subtypes.