Automatic speech recognition, especially speaker independent continuous speech recognition, isrncharacterized by great variability in word pronunciation, including many variants that differrngrossly from dictionary prototypes. This is one factor in the poor performance of automaticrnspeech recognizers on speaker independent speech recognition.rnOne approach to handling this variation consists of expanding the dictionary with phoneticrnsubstitution, insertion, and deletion rules. This study aimed at modeling multiple pronunciationsrnin speaker independent continuous speech recognition for Afaan Oromoo to handlernpronunciation variation. Hidden Markov Model and the Hidden Markov Modeling Toolkit werernused to implement it. For developing model for the language under study, a corpus containingrn754 sentences collected from Bariisaa news paper and Afaan Oromoo bible (New Testament)rnwas used. The data collected was preprocessed in line with the requirements of HTK. Phonemesrnwere taken as base unit for recognition. Knowledge based pronunciation variation modelingrntechnique was used for modeling words with multiple pronunciations. Thus two models wererndeveloped; one with canonical pronunciation and the other with alternate pronunciation torncompare their relative performance.rnAccordingly, the performance achieved using canonical pronunciation was 81.09% and 83.82 %rncorrect for sentences and words respectively with word accuracy of 80.91 % while thernperformance of alternate pronunciation was 83 .08% and 85. 11% sentences and words correctrnrespectively with 82.52% word accuracy.