To assess risk factors for treatment delay among NSCLC patients in Louisiana. Linked data between Research for Action Network (REACHnet) electronic health records and Louisiana Tumor Registry (LTR) was analyzed. The primary outcome was a dichotomous variable indicating treatment delay among NSCLC patients, defined as having longer than 6 weeks from diagnosis date to treatment initiation. Independent variables included age at diagnosis, sex, race, primary payer at diagnosis, census tract level, Charlson comorbidity index (CCI) score, smoking status, AJCC stage, other primary cancer diagnosis, and year of diagnosis. Classification supervised machine learning models including random forest, neural network, and logistic regression were used. Area under curve (AUC) was compared to choose the final model. 1,179 patients diagnosed with NSCLC from 2013 to 2017 were included. 75% of the entire dataset was used for training, the rest of the dataset was used for testing. CCI score, primary payer at diagnosis, smoking status at diagnosis, and AJCC stage at diagnosis were found to be important to the outcome through variable selection. Logistic regression had the highest AUC and was selected as the final model. After controlling for other covariates, Black patients were 1.755 times more likely to experience a treatment delay compared with White patients (p<0.01). Current smokers and former smokers were 2.566 (p<0.01) and 1.739 (p=0.09) times more likely to experience treatment delay compared with never smokers, respectively. Compared with patients diagnosed at Stage I, patients at Stage II were 1.657 times more likely to have treatment delay (p=0.07), at Stage III were about half as likely (p<0.01), and at Stage IV only 20% as likely to experience treatment delay (p<0.0001). Among patients with primary NSCLC in Louisiana diagnosed from 2013 to 2017, Black patients, smokers, and those diagnosed at early stages have higher probability of treatment delay.