Current prognosis in soft tissue sarcoma (STS) is analyzed and determined via the American Joint Commission on Cancer (AJCC) staging system, which incorporates tumor, node, metastasis framework, tumor size, depth, grade, nodal involvement, and presence of metastases to place patients into prognostic groups. This system strongly relies on tumor size categories T1-T4 with little account for anatomic depth of the tumor, even though these aspects have little impact on survival predictions. Recently, there has been a push to make prognostic models dependent on nomograms in an attempt to make prognostic predictions more personalized for patients. As a result, these models outperform the AJCC system. However, due to the lack of diversified input into these systems, there is concern that the accuracy of this model will be drastically different considering the clinical setting. Additionally, when these systems are trained on missing or incorrect data, their accuracy is also variable. In response to the lack of prognostic machine learning models that can accurately predict 2-5 year prognosis, this study produced a model with the intent of bridging this gap. The model they used was trained on the SEER database, which accounts for 28% of cancer cases in the United States. The model from this study showed higher accuracy compared to the conventional machine learning model for both 2 and 5 year predictions (AUC of 0.81 and 0.82), making it a beneficial tool to be integrated into prognosis consideration.