This paper summarizes the results from a series of large scale cyclic model load tests on geocell reinforced and unreinforced homogeneous sand subgrades. The sand subgrades are prepared at 70 % relative density thru pluviation technic in a large steel tank measuring 1 m3. A cyclic load of 0.7 and 7 kN was applied on a circular steel plate of 150 mm diameter to replicate a portion of load transfer from an equivalent single axle wheel load (ESAL) on the sand subgrade. The cyclic load was applied through a double acting dynamic hydraulic 100 kN actuator. A series of tests were conducted to study the effects of width and height of a geocell reinforcement on the cyclic behavior of sand subgrades. The performance improvement is quantified in terms of traffic benefit ratios (TBR), cumulative plastic deformations (CPD), modulus of subgrade reaction (ks), resilient modulus (Mr) and rut depth reduction (RDR) for different number of cycles and plate settlements. The configuration of the geocell that provides highest performance in terms of TBR, CPD, and RDR, is considered to be the optimum size of the geocell. It is observed that the optimum geometry of the geocell is capable of reducing the rut depth of about 75 % at 100 cycles and provide about a TBR of 23 at 10 % of settlement ratio. © 2014, Indian Geotechnical Society.