ENInflection of Lithuanian words should be modeled when creating computer programs for automatic stressing, lemmatizing etc. According to the inflexion method, all words of the Lithuanian language can be divided into noun-adjectives, verbs and non-inflectional words. Three separate modules having its own databases of morphemes and algorithms of inflection can be built based on these three groups. Some grammatical word forms have features appropriate to several groups; therefore, they can be modeled in several alternative ways. The following alternatives were investigated in this work: verbal nouns and participles assigned to the group of verbs vs. noun-adjectives, the nouns of the fifth declension (nominative case, singular) assigned to the group of non-inflectional words vs. noun-adjectives. Besides, the grouping of declinable parts of speech was investigated. Alternative methods were compared taking into account processor's time, memory and similarity of the model to the grammar of Lithuanian. Experiments showed small influence of modifications made to the speed and size of the algorithm.