The T-spherical fuzzy set is a useful tool to model evaluation values in decision-making process. To solve multiple attribute group decision-making problem with T-spherical fuzzy evaluation values and unknown attribute weights, this paper proposed T-Spherical fuzzy ORESTE method based on T-spherical fuzzy cross-entropy measures. First, some T-spherical fuzzy cross-entropy measures are developed, which are used to determine experts weights. Then, the T-Spherical fuzzy ORESTE method is introduced based on new cross-entropy measures, in which the evaluation values of alternatives and attributes are all in the form of T-spherical fuzzy values. Finally, a case of selecting transformation alternatives for old communities is presented to illustrate feasibility of the proposed method. Some comparisons of the proposed method with other T-Spherical fuzzy methods are also conducted to show effectiveness and reliability of the new method.