The objective of this research is to know the effectiveness of minimal pairs practice for teaching pronunciation to the first grade students 2006/2007 at SMPN 207 Meruya, West Jakarta. The research was conducted from August to September 2006 at SMP Negeri 207 Meruya, West Jakarta. The method of the research is pre-experiment. Pronunciation test is used for the instrument of the research. The technique of data analysis of this research is t-test analysis. The population of this research was 280 students which were distributed into seven classes and the writer took one class consisting of 37 students as the samples of this research employing random cluster sampling technique. Based on the analysis, the writer gets a result that the computed t obtained 15.79, the t distribution with the opportunity 0.95 and degree of freedom 36, the value of table t: 1.70. It means that to (15.79) > tt (1.70), so the null hypothesis is rejected and the alternative hypothesis is accepted. Therefore, it indicates that the teaching of pronunciation by using minimal pairs practice for the first grade students is effective. Since this is a pre-experiment, the conclusion needs to be experimented.