Unstable sitting is used commonly to evaluate trunk postural control (TPC), typically via measures based on center-of-pressure (CoP) time series. However, these measures do not directly reflect underlying control/movement strategies. We quantified trunk-pelvis coordination during unstable sitting using vector coding (VC) and correlated such coordination with CoP-based outcomes across varying task demands. Thirteen uninjured individuals (11 male/2 female) sat on an unstable chair at four instability levels, in a random order, defined relative to the individual gravitational gradient (∇G): 100, 75, 60, and 45%∇G. VC assessed trunk-pelvic coordination, and coupling angles classified movements as: 1) anti-phase, 2) in-phase, 3) trunk-phase, or 4) pelvic-phase. With decreasing %∇G (i.e., increasing instability), we found: increased anti-phase movement in the sagittal and frontal planes; decreased in-phase movement in the sagittal and frontal planes; and increased in-phase and pelvic-phase movement in the transverse plane. In the sagittal and frontal planes, we observed significant weak-to-moderate correlations between anti-phase and in-phase movements (0.288 < |ρ| < 0.549). Correlations between CoP-based measures and pelvic-phase and trunk-phase movements were typically weak and/or non-significant (|ρ| < 0.318). VC techniques discriminated between levels of instability during unstable sitting, identifying in-phase coordination (stiffening strategy) at lower instability levels and anti-phase coordination at higher instability levels. Compared to CoP-based measures, trunk coordination outcomes during unstable sitting provide measures of TPC that more directly quantify underlying movement strategies. These results can also serve as a baseline for future work investigating populations with impaired TPC (e.g., individuals with low back pain or limb loss).