This research attemps to propose an approache to make deciscions under uncertainity forrndesigning a dialogue manager for Amharic spoken dialogue system, Amharic as under-resourcedrnlanguage. A prototype Amharic Spoken Dialogue System was implemented on hotel andrnrestaurant address information domain for experimentation. Data for this research was collectedrnthrough a method called Wizard of Oz and domain knowledge is prepared using address searchrnwebsites. How to design a dialogue manager which is robust for languages especially with lowrnperforming Automatic Speech Recognition unit is a fundamental question of this research.rnPrevious studies on designing of spoken dialogue system for under resourced languages focusedrnmainly on Automatic Speech Recognizer. We reviewed methods and frameworks on dialoguernmanagement. Design of Partially Observable Markov Decision Processes (POMDP) basedrndialogue manager, which provides a principled framework to plan under uncertainity, yieldsrnrobustness. Low perfoming Spoken Dilaougue System (SDS) componenets, especially thernAutomatic Speech Recognizer (ASR) were considered the causes of uncertainity. Maintainingrnmultiple hypotheses (evidences) improves the correctness of the dialogue manager. Wernconducted experiments to test correctness score, error rate and robustness. With a maximum of 6rnN-best list and 20 partitions the correctness score grow by 14.19%. Increasing the number of nbestrnlist of 6 reduced the error rate by 5.78% and with 6 n-best list. The belief updated belowrn0.12 seconds with 20 partitions and 6-nbest list. And the dialogue manager was able to completerna task with an average 8.75 turns by 50% Word Error Rate. The finding from the researchrnillustrates that POMDP-based design approach to dialogue management is robust and possible torndevelop an improving spoken dialogue system for under-resourced languages