The KZFP/KAP1 (Krüppel associated box zinc finger proteins/KRAB-associated protein 1) complex is a transcriptional repressor that triggers the deposition of heterochromatin (H3K9me3) and DNA methylation. The main targets of the KZFP/KAP1 complex are transposable elements (TEs), which are generally maintained transcriptionally inactive to avoid retrotransposition and preserve genomic stability, and imprinting control regions (ICRs), where DNA methylation is protected to ensure proper parent-of-origin specific expression of crucial developmental genes. The role of the KZFP/KAP1 system in forming a heterochromatic environment is of particular relevance during the wave of genome-wide epigenetic reprogramming that takes place in the early embryo between fertilization and implantation. During these stages, TEs and ICRs are amongst the few sequences that retain the epigenetic information inherited from the gametes. While the importance of KZFP/KAP1 in maintenance of imprinting and TE silencing is well established, the mechanisms by which KAP1 counteracts demethylation and interacts with this epigenetically unstable environment to regulate TEs are still largely unknown. In this work, we uncover the complex interplay between KAP1, DNA methylation and active demethylation in maintaining imprinting and regulating TE expression in naïve embryonic stem cells. We first confirmed that KAP1 was able to bind and repress TEs in absence of DNA methylation. We then demonstrated that the KZFP/KAP1 complex was required to maintain heterochromatin and DNA methylation at ICRs and TEs, and was able to protect these sequences from active demethylation. We comprehensively characterized the responsiveness of TEs to removal of KAP1 in absence of DNA methylation and active demethylation, which unveiled a complex, multi-layered and integrant-specific regulation for these elements. ZFP57 is known to recruit KAP1 on ICRs, but here we identified ZNF445 as a second KZFP that specifically binds ICRs both in human and mouse. We found that ZNF445 regulates imprinting differentially in human versus murine embryonic stem cells. Our findings suggest for the first time that two different KZFPs could have evolved with species-specific roles in imprinting regulation. The high degree of conservation of ZNF445 in the human population argues in favour of a previously unsuspected critical role for this protein in development. This work provides useful insights in the role of KAP1 and its KZFPs partners both in preserving epigenetic memory and in transposon silencing during early development.