A statistical translation model is a mathematical model in which the process of
human language translation is statistically modeled. Model parameters are
automatically estimated using a corpus of translation pairs. This paper presents a
statistical model for translating the different kinds of English sentences into Bangla. A
lot of alignments are possible between translation of English and Bangla sentences. In
this paper, we have used a fertility model that finds the possible meanings of English
word into Bangla word with corresponding probabilities. For calculating the
probabilities of all possible alignments we have used distortion model and chosen the
best one that have the highest probability. For searching process, an efficient algorithm
called A* search algorithm has been used in this paper that provides a good result. We
have tested our proposed model using several English sentences with different word
lengths and got successful translation for most of the test cases.