Results of research on the correlation between pollutant emission rates and the engine operation states that determine the pollutant emissions have been presented. The tests were carried out on a compression ignition (CI) engine Cummins 6C8.3 in the NRTC (Non-Road Transient Cycle) test conditions. At the tests, the speed, torque, and effective power of the engine under test were chosen as the main characteristics of the engine operation states that determine pollutant emissions. Correlation dependences of pollutant emission rates on engine operation states have been shown. Mathematic models of pollutant emission rates have been defined as second-degree polynomial functions of engine speed and torque. To analyse the correlation between the sets under investigation, the Pearson’s linear correlation, Spearman’s rank correlation, Kruskal’s gamma correlation, and Kendall’s tau correlation theories were used. It has been found statistically justifiable to treat the examined pairs of sets of physical quantities as strongly correlated. Moreover, the quantities defining the engine operation state have been found to have a similar impact on the carbon monoxide and hydrocarbons emission rates, while their impact on the nitrogen oxides emission rates has been found to be quite different. For the carbon monoxide and hydrocarbons emission rates, the engine speed has been found to be the factor of the strongest impact; as regards the nitrogen oxides emission rates, they were most strongly affected by the torque and effective power of the engine.