Phenotypic records have gained in depth and complexity in the field of forest tree genetics and genomics, thanks to cumulated historical data and the raising of new high-throughput screening techniques.This leads to highly structured data such as multivariate, longitudinal, genomic or multi site. Their interplay with complex effects can quickly become overwhelming and difficult to handle.The breedR (Muñoz and Sanchez 2016) package provides frequentist and Bayesian statistical tools to build and fit a large variety of models, while taking care of most of the technical details. It is able to handle effects that are useful in forest trials such as spatial autocorrelation, competition, permanent environments, genotype-by-environment interaction, and environmentally-based plasticity, to cite just a few.breedR implements all models as a Linear Mixed Model, builds all the required incidence and structure matrices and relies on the BLUPF90 (Misztal 1999) suite of FORTRAN programs, providing reliable REML estimation or Gibbs sampling.In addition, it retrieves the results into convenient Robjects for subsequent inspection, diagnosis or plotting. BreedR is undergoing active development as part of the Trees4Future and ProCoGen projects under a Free and Open Source license (GPL-3), and backed by a rapidly growing community. We illustrate some of its capabilities using a case study on spatial analyses.