Nowadays, many learning applications using ASR system. ASR can capture children's interest and engage them in their learning (Husniza. H, Fauziah. AR, Sobihatun. NAS, 2012). ASR can also improve the quality of learning and teaching helps ensure that online learning is accessible to all through the cost-effective production of synchronized and captioned multimedia content (Mike. W., 2002) . So, IMELDA is one of the applications using ASR technologies to help young children in primary school in Malaysia. IMELDA encouraged children to take an interest in learning English. However, when testing IMELDA at school with six primary school children, limitations were noted in this ASR system that affect accuracy. In order to ensure the proper functioning of this application using ASR technology, it must be able to handle all possible effects on the performance of an ASR engine (Victoria. Y., 2012). The theory literature has examined the relationship between ASR recognition performance as measured. by precision where the % of the word correct, divided by the total number of words used. ASR ACCURACY ASR accuracy is a difficult problem for automatic speech recognition system to handle. This difficulty is due to a certain factor. According to Victoria Y (2012), these errors are caused by two factors: external and internal factors. An external factor which is the sound environment and an internal factor due to the error of the components and the language model (LM) by the ASR system. In IMELDA, the factor causes ASR problems both (e.g. child's voice, sound environment, pronunciation error and language model (LM) used in IMELDA which is not suitable for L2 .Sound environmentNoise in the environment is one of the external factors that determine... ... middle of paper ...... error occurs when younger children may not have correct pronunciation. Additionally, the model of an ASR motor in IMELDA is not suitable for L2 children because the phonetics of L2 children are different from those of L1 children Sometimes young children do not know how to articulate specific phonemes (Schotz, 2001). .The speech recognition system almost develops using the English language. So, the ASR system poses problems for L2 children, especially in terms of accuracy, because the pronunciation style of the English language is different from that of children. L1. According to Muhirwe.J (2005), to build a speech engine, we need a speech corpus which can be obtained from text and speech collections and both are used on the basis of statistical processing of speech processing. natural language (NLP). Cole et al. (1994) also state that the development of a speech corpus could involve the collection and transcription of data..