Finite element (FE) model updating is a framework for identification of FE model and its parameters consistent with the system measurements. For updating FE model with uncertain model parameters, the Bayesian model updating approach is widely adopted which uses probability logic for uncertainty quantification. This approach using modal data is conducted by mode matching or without mode matching approaches. Mode matching approach requires solving the eigenvalue problem and relative weighting factors for modal frequencies and mode shapes. Whereas, without mode matching approach adopts the eigenvalue equation instead of solving it and does not require weighting factors. Due to limitations in gathering data from the structure in a single setup, model updating can be performed using data from multiple setups. Sampling techniques are commonly adopted for generating samples from posterior distribution of model parameters and its statistics. Among various sampling techniques, Gibbs sampling is appropriate for multi-dimensional problems when conditional distributions are possible for the parameters. The present work aims to perform FE model updating using modal data from multiple setups. The effectiveness and efficiency of a modified Gibbs sampling method in Bayesian model updating are illustrated by 10 degrees of freedom (DOF) numerically simulated freedom (DOF) numerically simulated example.