UCL Machine Learning Repository is a source for quite a large number of datasets for different problems.
For what you are looking, change the default task to Regression. Then to find datasets with higher number of predictors, you can just find yourself, or can sort by the # Attributes column.
Hope this helps.