Objective To evaluate the accuracy of an intelligent prenatal risk assessment system.

Methods A random sample of 1 387 pregnant women from a maternal and child health institution was assessed using both manual and intelligent prenatal risk assessment methods. The effectiveness of the intelligent prenatal risk assessment system was evaluated by comparing its authenticity and reliability with manual assessments.

Results Compared to manual prenatal risk assessment, the intelligent system’s sensitivity for the assessment of pregnant women categorized by green (low risk), yellow (general risk), orange (relatively high risk), red (high risk), and purple (pregnant women with infectious diseases) was 100.0%, 98.0%, 98.2%, 99.5%, and 95.1%, respectively. The specificity was 98.9%, 99.2%, 99.9%, 99.9%, and 100.0%, respectively. The Youden’s J indices were 0.989, 0.972, 0.981, 0.994, and 0.951, respectively.

Conclusion The intelligent prenatal risk assessment system achieves a relatively accurate assessment outcome. By ensuring the efficiency and accuracy of risk assessment while significantly reducing human resource costs, the system has a high potential for widespread adoption.