This paper presents an evaluation of several approaches of plantsspecies distribution modeling based on spatial, environmental and co-occurrences data using machine learning methods. In particular, we re-evaluate the environmental convolutional neural network model that ob-tained the best performance of the GeoLifeCLEF 2018 challenge but on arevised dataset that fixes some of the issues of the previous one. We alsogo deeper in the analysis of co-occurrences information by evaluating anew model that jointly takes environmental variables and co-occurrencesas inputs of an end-to-end network. Results show that the environmentalmodels are the best performing methods and that there is a significantamount of complementary information between co-occurrences and envi-ronment. Indeed, the model learned on both inputs allows a significant performance gain compared to the environmental model alone.