Analysis of the association between antibodies against bovine leukemia virus (BLV), BLV proviral load, and white blood cell (WBC) and lymphocyte counts was performed with 774 dairy cows. The average age, WBC counts and lymphoid cell counts tended to be higher in BLV antibody-positive cows than in antibody-negative cows. There was a similar trend in levels of proviral DNA. We analyzed age, WBC counts and lymphocyte counts by principal component analyses to create a distribution chart of the principle component scores. Using the chart, we categorized cows into four quadrants based on additional information, such as the presence of antibody and the levels of proviral DNA. Antibody-positive cows and cows with high BLV proviral load were found mostly in one quadrant of the chart, indicating that it is possible to predict the risk of infection without any knowledge on antibody status by using information, such as WBC counts as a biomarker. When only antibody-positive cows were included in the analysis, a characteristic distribution of different levels of proviral DNA was seen in the quadrants, suggesting that it is possible to estimate the extent of bovine leukosis infection by using this analysis. For this analysis and categorization of the cows into quadrants, we computed a mathematical formulation using discriminant analysis based on age and WBC and lymphocyte counts. This mathematical formulation for the hematological preliminary diagnosis of the disease is recommended as a screening tool to monitor bovine leukosis.