The study was aimed to examine the effectiveness of the English conversation to improve students' speaking skill focusing on whether English conversation effective in speaking skill and how is the effectiveness of the English conversation to improve students' speaking skill. The research was designed in a pre-experimental pre-test and post-test research, where the research only applied the English conversation to the experimental group that consisted of 28 students. The sample of the study was taken by using cluster random sampling technique. It was taken randomly of 148 second grader students of MTs. Negeri Masbagik as the population of the study. The gaining data of pre-test and post-test was taken by requiring the students to make and practice a conversation in front of the class based on the material that was learnt before. Later, the pre-test and post-test data was analyzed by using descriptive statistics and Paired Sample t-test to test the hypothesis. Having collected and calculated the data by using descriptive statistics, the present researcher found out that descriptive statistic result showed the mean score of the pre-test was 7.21 while the mean of the post-test was 9.25. The result of hypothesis testing by using Paired-Sample t-test was t(27) = 11.68 at p = .000. It was lower than .05 so it means that the null hypothesis (Ho) was rejected and the alternative hypothesis (Ha) was accepted. In other word that the English conversation was significantly effective in speaking skill.