This study addresses the challenge of processing the complex morphology of Ge’ez, an ancient southeast Semitic liturgical language. The research develops a comprehensive finite-state morphological analyzer and generator for all Ge’ez verb categories using bidirectional finite-state technology. The complexity of Ge’ez’s non-concatenative morphology, where consonantal roots receive vowel patterns through interdigitation and the absence of native speakers, presents unique challenges for computational processing. The study implements a rule-based analyzer using Foma’s finite-state framework and adopts the washära classification system, which recognizes eight head verbs. The morphological analyzer integrates finite-state transducers with lexc-based lexicon development, incorporating roots, affixes, vowel intercalation rules, and morphological alternations. For evaluation, a gold-standard dataset comprising 1365 verbs was compiled from the Ge’ez Bible and prayer book, with manual annotation by Ge’ez experts. The analyzer achieved an accuracy of 97.29% and a precision of 80.24% when evaluated against the gold-standard dataset, demonstrating significant improvement over previous approaches. Compared to earlier studies that focused on single verb categories or achieved limited accuracy, this analyzer successfully processes all verb types, including irregular verbs, and provides analysis and generation capabilities. This tool establishes a foundation for developing advanced NLP applications in Ge’ez, including machine translation and lexicography.