The authors of this paper applied a new approach combining text mining and principal component analysis (PCA) to objectively determine the actual state of regional COVID-19 strategy meetings and verified its utility. The authors used text mining to analyze meeting minutes and extracted words with high phase ubiquity by co-occurrence analysis. Then, they selected words symbolizing the meeting contents ("report," "prevention," "rules," and "decision") and performed PCA using the occurrence rates of these words as variables. Two principal components (PC1, PC2) were set. For PC1, we observed maximum factor loading for "decision" (0.81) and minimum for "report" (-0.72), so we considered this axis to show the "depth of meeting discussions." For PC2, we observed maximum factor loading for "prevention" (0.81) and minimum for "rule" (-0.76). We considered this axis to show "regional infection status." When we created a plot of all 44 meetings, Phase 1 occurred in quadrants 3 to 4 (knowledge sharing), phase 2 began in quadrant 1 (preparation for spread), and phase 3 shifted to quadrant 2 (response to spread) with significant differences between these phases. Our findings suggest that the actual state of regional COVID-19 strategy meetings could be objectively determined by using a combination of text mining and PCA.