Metamaterials with extreme thermoelastic properties (i.e. thermal expansion and stress) offer significant value in applications that require maintaining thermal geometric stability, e.g., aerospace equipment. However, material uncertainty has a great impact on the thermoelastic properties of such multiphase composites. Thus, a robust topology optimization design for the thermoelastic composites under hybrid uncertainties of material property is investigated. The distribution of base materials in the periodic unit cell is achieved by multi-phase solid isotropic material with penalization method and the equivalent properties of composites are evaluated by the numerical homogenization method. To obtain a clear material interfaces in the unit cell, a filtering method named the element perimeter control method is proposed. Considering the uncertainties of material property, the linear thermal expansion coefficients of two constituent materials are assumed to be a random variable and an interval variable, respectively. A robust topology optimization model based on Polynomial-Chaos-Chebyshev-Interval method is proposed, where the Polynomial chaos and Chebyshev interval functions are integrated to perform uncertainty analysis. The robust objective function is formulated by a combination of average mean, average and bandwidth of standard deviation of the objective value. The effectiveness of the proposed robust topology optimization methodology is illustrated with several numerical examples.