Background. Many studies have reported on the surgical outcomes of soft tissue sarcoma. However, there was no longitudinal cohort study. Because time is the most valuable factor for functional recovery, adjusting time value was the key for finding the causal relationship between other risk factors and postoperative function. Therefore, existing cross-sectional studies can neither fully explain the causal relationship between the risk factors and the functional score nor predict functional recovery. The aim of this study was to determine important predictive factors that affect postoperative functional outcomes and longitudinal changes in functional outcomes in patients who had undergone limb-sparing surgery (LSS) for soft tissue sarcoma (STS). Methods. Between January 2008 and December 2014, we retrospectively enrolled 150 patients who had undergone LSS for STS and had been assessed for postoperative functional outcomes with questionnaires. To evaluate functional outcomes, we used the Musculoskeletal Tumor Society (MSTS) score and Toronto Extremity Salvage Score (TESS). Multivariate generalized estimating equation (GEE) analysis was used to identify the predictive factors, including size, stage, and anatomic location of tumor, bone resection, flap reconstruction, age, and time after surgery. Each continuous variable such as age and time after surgery was explored for statistically significant cutoff points using the Wilcoxon rank sum test. Results. Functional scores significantly improved until the second year after surgery and plateaued for the rest of the 5-year period. Age p<0.0001, bone resection p=0.0004, and time after surgery p<0.0001 were identified as significant predictive factors. The functional score was significantly higher in patients younger than 47 years old. Conclusions. Functional outcomes can improve until the second year after surgery. Patients who were older than 47 and underwent bone resection may have poor final functional outcomes.