The article examines criteria for the typological classification of political texts and the main problems arising in Uzbek-English translation. The empirical material consists of 6,428 parallel segments from an electronic corpus compiled from official https://lex.uz/ sources, including presidential resolutions, strategic programs, administrative and legal documents, and international treaties. Political texts are classified according to institutional status, communicative purpose, dominant function, genre, syntactic organization and pragmatic orientation. The proposed taxonomy distinguishes normative-legal, program-strategic, administrative-institutional, diplomatic, socio-political and hybrid text types. Comparative-typological, functional-pragmatic, discourse and corpus analyses show that the major translation difficulties concern terminological equivalence, deontic modality, performative formulae, institutional names, complex nominal structures, the representation of agency and ideologically or pragmatically marked meanings. Examples from the official corpus demonstrate that literal matching does not necessarily ensure functional or legal equivalence. The study proposes a multi-stage translation model based on the preliminary identification of text type, terminology standardization, harmonization of modal and performative patterns, and quality control through parallel-corpus comparison. The model can support the preparation, editing and evaluation of official English translations of Uzbek political and legal documents.