The poor and inconsistent physicochemical properties of bio-oil are inhibiting its industrialized production. We investigated the variability in properties of switchgrass bio-oil produced at three pyrolysis temperatures (T=450, 500, and 550 degrees C) and three feedstock moisture contents (MC=5%, 10%, and 15%) in a 3x3 factorial experiment in order to exploit opportunities to improve bio-oil properties through optimization of pyrolysis parameters. Results showed that even with the single type of feedstock and pyrolysis system, the two main factors and their interaction caused large variations in bio-oil yield and most of the measured physicochemical properties. Following improvements of bio-oil properties could be individually achieved by selecting an optimal pyrolysis condition (shown in parenthesis) comparing with the worst case: increase of bio-oil yield by more than twofold (MC=10%, T=450 degrees C), increase of pH by 20.4% from 2.74 to 3.3 (MC=10%, T=550 degrees C), increase of higher heating value by 18.1% from 16.6 to 19.6 MJ/kg (MC=10%, T=450 degrees C), decrease of density by 5.9% from 1.18 to 1.11 g/cm(3) (MC=5%, T=550 degrees C), decrease of water content by 36% from 31.4 to 20.1 wt.% (MC=5%, T=450 degrees C), decrease of viscosity by 40% from 28.2 to 17 centistokes (MC=5%, T=550 degrees C), decrease of solid content by 57% from 2.86 to 1.23 wt.% (MC=15%, T=550 degrees C), and decrease of ash content by 41.9% from 0.62 to 0.36 wt.% (MC=15%, T=550 degrees C). There is no single, clear-cut optimal condition that can satisfy the criteria for a bio-oil product with all the desired properties. Trade-offs should be balanced according to the usage of the end-products.