Existing studies on the health effects of smoke fine particulate matters (PM), a primary emission from wildfires, have often lacked comparison with other air pollutants, focused primarily on acute exposures, and not applied causal methods. In the present study, we obtained county-level, three-year average cardiovascular hospitalization rates for Medicare beneficiaries across the contiguous US between 2006-2016 from the Centers for Disease Control and Prevention. These data were linked with spatio-temporal estimates of smoke PM, non-smoke PM, nitrogen dioxide (NO), ozone, and county-level confounders. We used a difference-in-differences method to evaluate causal effects of three-year moving average exposures (lag 0-2, 1-3, 2-4, or 3-5 year) to the four pollutants on hospitalization rates for total cardiovascular disease (CVD) and its two major subtypes: heart disease and stroke. We found that, for total CVD, the absolute change in hospitalization rate associated with smoke PM increased with longer lag periods: from -0.879 (95% confidence interval [CI]: -2.528, 0.771) at lag 0-2 to 7.538 (95% CI: 4.594, 10.481) at lag 3-5 per 1 μg/m increase in exposure per 1,000 people. The effect of non-smoke PM was smaller and diminished over time. NO and ozone had even smaller effects per 1 part per billion increases in exposure. Similar patterns were seen for heart disease. For stroke, all pollutants had minimal and mostly non-significant effects. More rural and lower-income counties experienced greater risks. These findings suggested the need to prioritize wildfire management in addition to traditional air quality control strategies.