This study investigated the emission of PM10 and PM2.5 (particulates with diameters of less than 10 µm and 2.5 µm, respectively) and the chemical composition of PM2.5 from laboratory combustion of five Australian vegetation types (three grasslands, a woodland and a forest). A mix of plants representative of Banksia (woodland) and Jarrah (forest) and three types of grasses (Spinifex – Triodia basedowii; Kimberley grass – Sehima nervosum and Heteropogon contortus; and an invasive grass (Veldt) – Ehrharta calycina) were burnt in 9 combustion conditions comprised of 3 fuel moisture levels (dry, moist, wet) and 3 air flow rates (no, low and high flow). PM (particulate matter) samples were collected onto filters and measured using gravimetric analysis. PM2.5 was then extracted and analyzed for water-soluble metals and polycyclic aromatic hydrocarbons (PAH) concentrations. The largest proportion of PM10 (98%) from vegetation fires was PM2.5. Banksia yielded the highest PM2.5 emission factor (EF), followed by Jarrah and Spinifex. Veldt grass combustion generated significantly higher emissions of PM2.5 compared with the other two grass types. High moisture contents and flow rates resulted in larger emissions of PM2.5. A strong correlation (R2 = 0.84) was observed between the EF for PM2.5 and combustion efficiency, suggesting higher PM emission with lower combustion efficiencies. Potassium and sodium were the most abundant PM2.5-bound water soluble metals, accounting for more than 97% of the total mass of metals analyzed. PAHs were found in significant concentrations, including the carcinogenic benzo(a)pyrene. Pyrene and fluoranthene were the most abundant PAHs detected, accounting for nearly 40% mass of the total PAHs. Indeno(1,2,3-cd)pyrene and benzo(g,h,i)perylene ratio (IND/IND + BghiP) appeared to be produced in a diagnostic ratio that indicated that the PAHs were derived from vegetation fires rather than other sources of emissions. The EF for PM2.5 and its chemical composition (water-soluble metals and PAHs) were strongly influenced by the type of vegetation burned. The results presented in this study could be useful in predicting the risks of human health effects on firefighters and the public who may be exposed to regular bushfires in Australia.