This study investigates the relationship between chemical analyses and magnetic susceptibility of zinnwaldite through magnetic separation of various size fractions. Statistical analyses were used to increase information about magnetic properties of this mineral as a future source of lithium. Statistical modeling indicated that magnetic susceptibility (as a main factor of magnetic separation) accurately can be predicted based on cations content of zinnwaldite. However the size of particles had a significant effect on magnetic susceptibility. The small difference between the estimated and measured values for the non-linear relationship of this prediction (less than 1 (10(-8) m(3)/kg)) shows that these accurate theoretical techniques can be also applied to estimate magnetic properties of zinnwaldite in other resources, and in-situ analysis. (C) 2014 Elsevier Ltd. All rights reserved.