Although subjective global assessment (SGA) is a widely used tool for nutritional investigation, the scores are dependent on the inspectors’ subjective opinions, and there are only few studies that directly assessed the usefulness of SGA and modified SGA in incident peritoneal dialysis (PD) patients. A total of 365 incident PD patients between 2009 and 2015 were enrolled and measured with SGA and calculated using serum albumin and total iron binding capacity (TIBC) levels for weighted SGA. Cox analyses were performed to delineate the association between SGA or weighted SGA and all-cause mortality, and a receiver-operating characteristic was conducted to reveal the additional benefit of weighted SGA on predicting adverse clinical outcomes. The Kaplan–Meier curve showed that the cumulative survival rate in patients with “Good nutrition” (G1) was significantly higher compared to those with “Mild to severe malnutrition” (G2). G2 was significantly associated with an increase in the mortality even after adjusting for several covariates compared with G1. Moreover, a 1-unit increase in weighted SGA was also significantly correlated with mortality after adjustment of the same covariates, while G2 was not significantly associated with an increase in the mortality among young-aged (under 65 years) groups. Meanwhile, a 1-unit increase in weighted SGA was significantly related to an increase in mortality in all the subgroup analyses. Furthermore, the AUCs of weighted SGAs in all groups were significantly increased compared with those of SGA alone. In conclusions, the evaluation of nutritional status based on SGA in incident PD patients might be useful for predicting mortality. However, weighted SGA with serum albumin and TIBC can provide additional predictive power for mortality compared with SGA alone in incident PD patients.