A novel local feature descriptor, called Local Contourlet Binary Pattern (LCBP), was developed in this paper. LCBP provides a multiscale and multidirectional representation for images since it integrates multiscale geometric analysis and local binary pattern operators. With the quadtree structure of LCBP and simplicity of the model itself, the LCBP coefficients were modeled by a two-state HMT that is in accordance with the intra-band, inter-band and inter-directional distributions of LCBP coefficients. Based on the LCBP-HMT model, an object classification method was further proposed to extract parameters of the LCBP-HMT model as features and classify the query samples by comparing the Kullback-Liebler distance between features of the query samples and that of the prototype objects. Experimental results illustrate the superiority of the LCBP over traditional wavelet features and Gaussian density function model features of contourlet coefficients in terms of the discrimination performance.