Objectives: To profile the learning styles of nursing students from the five regions of Brazil and one region of the United States. In addition, specific objectives were: to verify the predominant learning style in beginning Nursing students, to identify the relationship between the learning styles and the sociodemographic and teaching variables, to compare the learning styles of the Nursing students from different regions, to measure the reliability version 3.1 of Kolb's Learning Styles Inventory and compare the results with the normative values of version 3.1 of Kolb's Learning Styles Inventory. Method: Observational, cross-sectional, correlational and quantitative study. The sample was formed by 176 students. Identification data, teaching preferences and academic background were collected, in addition to the application of version 3.1 of the Learning Style Inventory, translated and validated for Brazilian Portuguese. Data collection was performed by the main researcher and assistant researchers, after training. Descriptive analysis was performed, followed by verification of data distribution using the Shapiro-Wilk test. Inferential statistics was applied for the measures of comparison, correlation and association between the variables, using uni or bilateral tests and checking the internal consistency of the inventory (Cronbach's alpha). Multiple linear regression models were constructed using generalized linear models, considering the inventory scores as dependent variables. Results: There was a predominance of students with a Divergers learning style (34.10%), while 28.41% were Assimilators, 23.86%, Accommodators and 13.63%, Convergers. In Brazilian centers located in the Southeast and North Regions, as well as in the USA, Assimilators students predominated. The average score of Concrete Experience for those who studied in a public school was higher (average = 26.22; p = 0.0019) and showed a difference in Brazilian state, federal and American universities (p = 0.0149). When comparing the results obtained with the normative values of version 3.1 of the inventory, it was observed that the results were similar, being significant when we verified the values of OR (Student's t test, p <.0001) and CA (Student's t test, p <.0001). The internal consistency of the instrument for the learning characteristics and its items was 0.