Four different models have been developed and tested using evolutionary polynomial regression method. Considering three clay types including intact, fissured and leda clays, three polynomial models have been presented for each one and a fourth one for all clay types using a wide range database of in-situ (CPT) and laboratory (oedometer) tests. Models have been tested using new data (not used in training stage) and comparing with target OCR values. Results were in good agreement with reference values from laboratory tests and have been compared with existing methods. Overall, models have good performance in comparison to available approaches. © 2014 ejge.