This study examined whether the degree of novel information processing at work (NPW) attenuates cognitive aging across 14 years for adults 50+ in the United States and how NPW links with job complexity. To answer these questions, we used data (N = 4,252) from the Health and Retirement Study. Detailed information on occupational characteristics from O*Net between 2000 and 2014 was used to assess NPW and matched with participants' occupational codes across time. Multilevel transition models were employed to estimate the relationship between NPW and cognitive functioning across time and to explore the moderating effect of cognitive level. Our results showed that exposure to more NPW across time attenuates cognitive decline as indicated by immediate word recall and serial 7s performance, while adjusting for baseline age, leisure, volunteering activities, cognition at previous wave, and other covariates. This buffering effect of NPW is reduced but sustained when controlling for change in job complexity, indicating that, as expected, NPW and job complexity are linked. We also found that particularly for workers with lower levels of cognitive performance, exposure to NPW across 14 years was linked with less decline in serial 7s performance (controlling for job complexity). This, however, was not the case for immediate word recall performance, which asks for further study. Overall, our findings suggest that mental stimulation through NPW is one mechanism (in the work context) that buffers cognitive decline.