Collaborative filtering (CF)-based recommenders work by estimating a user's potential preferences on unobserved items referring to the other users' observed preferences. Slope-one, as a well-known CF recommender, is widely adopted in industrial applications owing to it's (a) competitive prediction accuracy for user's potential preferences, (b) high computational efficiency, and (c) ease of implementation. However, current Slope-one-based algorithms are all designed for static datasets, which are contradictory to real situations where dynamic datasets are mostly involved. This paper focuses on designing incremental Slope-one recommenders able to address dynamic datasets, reflecting their variations instantly without retraining the whole model. To do so, we have carefully analyzed the parameter training processing of Slope-one-based recommenders to design the incremental update rules for involved parameters reflecting data increments in dynamic environments. Three incremental Slope-one recommenders, including the incremental Slope-one, incremental weighted Slope-one, and incremental bi-polar slope one, are proposed. Experimental results on two large real datasets indicate that the proposed incremental slope-one recommenders can correctly reflect the increments of dynamic datasets with high computational efficiency. (C) 2017 Elsevier B.V. All rights reserved.