BACKGROUND: Non-communicable diseases (NCDs) are the world's leading cause of death and disability. Global implementation of WHO-recommended NCD policies has been increasing with time, but in 2019 fewer than half of these policies had been implemented globally. In 2022, WHO released updated data on NCD policy implementation, on the basis of surveys conducted in 2021 during the COVID-19 pandemic. We sought to examine whether the trajectory of global policy implementation changed during this period. METHODS: In this repeated cross-sectional analysis, we used data from the 2015, 2017, 2020, and 2022 WHO progress monitors to calculate NCD policy implementation scores for all 194 WHO member states. We used Welch's ANOVA and Games-Howell post-hoc pairwise testing to examine changes in mean implementation scores for 19 WHO-recommended NCD policies, with assessment at the global, geographical, geopolitical, and country-income levels. We collated sales data on tobacco, alcohol, and junk foods to examine the association between changes in sales and the predicted probability of implementation of policies targeting these products. We also calculated the Corporate Financial Influence Index (CFII) for each country, which was used to assess the association between corporate influence and policy implementation. We used logistic regression to assess the relationship between product sales and the probability of implementing related policies. The relationship between CFII and policy implementation was assessed with Pearson's correlation analysis and random-effects multivariate regression. FINDINGS: Across the 194 countries, in the years preceding publication of each progress monitor, mean total policy implementation score (out of a potential 18·0) was 7·0 (SD 3·5) in 2014, 8·2 (3·5) in 2016, 8·6 (3·6) in 2019, and 8·6 (3·6) in 2021. Only the differences in mean implementation score between 2014 and the other three report years were deemed statistically significant (pairwise p<0·05). Thus the steady improvement in mean global NCD policy implementation stalled in 2021 at 47·8%. However, from 2019 to 2021, we identified shifts in individual policies: global mean implementation scores increased for policies on tobacco, clinical guidelines, salt, and child food marketing, and decreased for policies on alcohol, breastmilk substitute marketing, physical activity mass media campaigns, risk factor surveys, and national NCD plans and targets. Six of the seven policies with the lowest levels of implementation (global mean score <0·4 out of a potential 1·0) in both 2019 and 2021 were related to tobacco, alcohol, and unhealthy food. From 2020 onwards, we identified weak or no associations between sales of tobacco, alcohol, and junk foods and the predicted probability of implementing policies related to each commodity. Country-level CFII was significantly associated with total policy implementation score (Pearson's r -0·49, 95% CI -0·59 to -0·36), and this finding was supported in multivariate modelling for all policies combined and for all commercial policies except alcohol policies. INTERPRETATION: NCD policy implementation has stagnated. Progress in the implementation of some policies is matched by decreased implementation of others, particularly those related to unhealthy commodities. To prevent NCDs and their consequences, and attain the Sustainable Development Goals, the rate of NCD policy adoption must be substantially and urgently increased before the next NCD progress monitor and UN high-level meeting on NCDs in 2024. FUNDING: None.