That depends in part on the complexity of your motif. If yout TF binds a simpel 6-8 bp sequence, such as a nuclear hormone receptor, then you cannot rely on an informatic approach that would generate thousands of hits. There is no substitute for actual experimentation. A ChIP-Seq approach would be best for bona fide targets, coupeld to a knock down and expression array or RNA-Seq.