The research aims to describe the speech variation of the Benteng Somba Opu community. The research method used is descriptive qualitative. Data were gaining from the verbal speech of the Benteng Somba Opu community in the form of dialogue. Data collection techniques conducted by doing observation and recording techniques. The data analysis carried out through data identification, data classification, data interpretation, and description of the result based on the sociolinguistics theory. The results of the research show that the Benteng Somba Opu community speech variation marked by using suffix, prefix, particle, and code-mixing. The suffix form contains -mi, -ma, -mo, -na, -nu, -ni, -ka, -ki, -ko, -ku, -ji, -pi, -di, -i, and - e. The prefix form marked by na-, ku-, ta-, and -i. The particle seems to the words, iyo, tong, tomma (i), mako, mami, paeng, dih, and tawwa. Code mixing conveyed using the English language (off and reward), and Arabic (alhamdulillah). Those variations are adjusting with the level of social, position, age, and familiarity between the speaker and the addressee in the Benteng Somba Opu community, resulting in a good, respectful, and polite speech.