Job deterioration and machine learning co-exist in various real life scheduling settings. This paper studies several single machine scheduling problems under the joint effect of nonlinear job deterioration and time-dependent learning. We assume that the processing time of a job increases when its processing is delayed. In addition, it is assumed that the machine undergoes a learning process, decreasing the time required to process a given job. The following objectives are considered: the makespan, the sum of completion times (square) and the maximum lateness. We derive polynomial-time optimal solutions for all the objectives. Crown Copyright (C) 2009 Published by Elsevier Ltd. All rights reserved.