This study established the association between executive directors’ remuneration and earnings management among banks in Kenya. The research adopted descriptive design. A sample of 34 banks operational over the 12-year period, 2007-2018 was selected. The study used semi structured questionnaires for primary data collection from a sample of accountants, internal auditors and finance managers operating in banks in Kenya. The study used secondary data from the individual financial statements and annual reports by commercial banks to collect data regarding information related to the total aggregate amount of executive directors’ remuneration and the control variables. Discretionary loan loss provision model was used to assess for cases of earnings management from the secondary data retrieved. Both the ordinary least square regression and fixed effects regression models were used. Primary data was analyzed using descriptive statistics and triangulated to observations from secondary data. The results retrieved from both analyses was used for comparison to establish whether there was an agreement or conflict of the findings between the primary and secondary data. The discretionary loan loss provision model found that there was income-increasing and income decreasing discretionary accruals, proxy for earnings management among the banks, which was consistent with the respondents’ beliefs. In addition, consistent with the agency theory hypothesis, the study established that there was negative but not statistically significant association between executive compensation and earnings management. However, respondents indicated that stock compensation and bonus payment had a positive impact on earnings management while cash compensation had none. Policy makers and regulators in the industry are encouraged to monitor patterns and trends of earnings management practices for guidelines issuance. Researchers are encouraged to explore other bank specific models of estimating discretionary accruals. One of the limitations was that some of the respondents were reluctant to take part in the study citing policies that prevented them from providing information before obtaining approvals.