Industrial accident insurance systems, especially in developing economies such as Kazakhstan, are hard-pressed to not only foresee work-related dangers accurately but also to make the best use of their resources. The said insurance system needs a thorough machine-learning technique to predict and reduce risks in the industry in such a way that it serves both employees’ welfare and the system’s financial stability. The study simulated a dataset of 5,000 enterprises representing 8 different industrial sectors in which 13 features of each company such as organization, safety investment, and risk exposure were carefully delineated, and a Random Forest classification model was applied. Various metrics illustrated the model’s impressive effectiveness, achieving 87.3% accuracy, 0.912 ROC-AUC score, and 0.867 weighted F1-score, thus providing a very good comparison with existing methods for occupational risk assessment. Feature importance analysis shows that the industry sector (28%), previous incidents (22%), and equipment age (18%) account for the most significant contributors to workplace accident risk. The evaluation of the financial implications of the policy shows the potential of real money to be saved annually to the tune of 843 million KZT over and above the 211 million KZT implementation costs, thus offering a good 4:1 return on investment and a time payback period of around 3 months. The risk-based premium optimization framework proposed facilitates fair insurance cost sharing by differentially pricing (30% discount for low-risk, 2.5× multiplier for high-risk enterprises) and at the same time providing a safety incentive. Performance of the model across various industries such as mining (89.2%), manufacturing (85.8%), construction (88.1%), and service industries (86.4%) makes implementation at the national level not only possible but also reliable. The present study is the first in Kazakhstan to fully integrate machine learning techniques with economic optimization while taking into consideration the country’s industrial context and hence serving as a methodological platform for the modernization of industrial accident insurance systems in other developing economies.