In the second language acquisition process, there are always some language errors from learners. This study aims to analyze errors in the use of compound sentence particles that express contradictions or known as gyakusetsu no setsuzokujoshi such as nagara, noni, temo, and kuseni toward Japanese language learners. To find out the types and causes of these errors, this study used a qualitative descriptive approach to analyze language errors. Data collection conducted in this research was obtained from 63 intermediate-level Japanese learners from Universitas Pendidikan Indonesia and STBA Yapari ABA Bandung. Based on the research results, each category of questions has types of errors that can be identified as overgeneralization, ignorance of rule restriction, incomplete application of rules, and false concept hypothesized. The factor causing the error was interference from Bahasa Indonesia as their mother tongue and learners' low distinction and understanding of the use of those setsuzokujoshi.