This work presents a linear regression model based on the number of registered business entities and the number of their employees. All counties of Republic of Croatia are covered. Data on the number of registered business entities and employees to be processed in this analysis refer to the status as of March 31, 2015. Business entities include legal persons, as well as crafts and free occupations. The number of employees in business entities is used to express the structure and size of business entities. This paper shows the relationship between the number of registered business entities and the number of employees in the same. Simple linear regression is used to measure the influence of an independent variable on the dependent variable, as shown in this paper on the example of employment in business entities, where the number of registered business entities is an independent variable and the number of employees is a dependent variable. The regression model used in the calculations, apart from the parameters a and b, also contains a random error that includes variables that are not included in the set model and affect the dependent variable. Using the least squares method, in the scatter diagram, the regression line is computed, where the total squared distance of all points x and y of the drawn regression line is minimal. The coefficient of determination calculated in the example of employment in business entities shows that it is acceptable to make prediction based on the regression model. The higher the coefficient of determination, the regression model is more representative. The coefficient of correlation shows a strong correlation between the number of registered business entities and the variables of the number of employees