Objective To determine the agreement between the cardiovascular disease (CVD) risk predictions computed with the WHO non-laboratory-based model and laboratory-based model in a nationally representative sample of Peruvian adults. Design Cross-sectional analysis of a national health survey. Methods Absolute CVD risk was computed with the 2019 WHO laboratory-based and non-laboratory-based models. The risk predictions from both models were compared with Bland-Altman plots, Lin's concordance coefficient correlation (LCCC), and kappa statistics, stratified by sex, age, body mass index categories, smoking and diabetes status. Results 663 people aged 30-59 years were included in the analysis. Overall, there were no substantial differences between the mean CVD risk computed with the laboratory-based model 2.0% (95% CI 1.8% to 2.2%) and the non-laboratory-based model 2.0% (95% CI 1.8% to 2.1%). In the Bland-Altman plots, the limits of agreement were the widest among people with diabetes (-0.21; 4.37) compared with people without diabetes (-1.17; 0.95). The lowest agreement as per the LCCC was also seen in people with diabetes (0.74 (95% CI 0.63 to 0.82)), the same was observed with the kappa statistic (kappa=0.36). In general, agreement between the scores was appropriate in terms of clinical significance. Conclusions The absolute cardiovascular predicted risk was similar between the laboratory-based and non-laboratory-based 2019 WHO cardiovascular risk models. Pending validation from longitudinal studies, the non-laboratory-based model (instead of the laboratory-based) could be used when assessing CVD risk in Peruvian population.