Objectives: A key policy question for the government of Uganda is how to equitably allocate primary health care pharmaceutical budgets to districts. This paper seeks to identify variables influencing current primary health care pharmaceutical expenditure and their usefulness in allocating prospective pharmaceutical budgets to districts. Methods: This was a cross sectional, retrospective observational study using secondary administrative data. We collected data on the value of pharmaceuticals procured by primary health care facilities in each district from National Medical Stores for the financial year 2011/2012. The dependent variable was expressed as per capita district pharmaceutical expenditure. By reviewing literature we identified 26 potential explanatory variables. They include supply, need and demand, and health system organization variables that may influence the demand and supply of health services and the corresponding pharmaceutical expenditure. We collected secondary data for these variables for all the districts in Uganda (n = 112). We performed econometric analysis to estimate parameters of various regression models. Results: There is a significant correlation between per capita district pharmaceutical expenditure and total district population, rural poverty, access to drinking water and outpatient department (OPD) per capita utilisation.(P<0.01). The percentage of health centre IIIs (HC III) among each district’s health facilities is significantly correlated with per capita pharmaceutical expenditure (P < 0.05). OPD per capita utilisation has a relatively strong correlation with per capita pharmaceutical expenditure (r = 0.498); all the other significant factors are weakly correlated with per capita pharmaceutical expenditure (r<0.5). From several iterations of an initially developed model, the proposed final model for explaining per capita pharmaceutical expenditure explains about 53% of the variation in pharmaceutical expenditure among districts in Uganda (Adjusted R2 = 0.528). All variables in the model are significant (p<0.01). Conclusions: From evaluation of the various models, proposed variables to consider in allocating prospective primary health care pharmaceutical budgets to districts in Uganda are: district outpatient department attendance per capita, total district population, total number of government health facilities in the district and the district human poverty index.