The magnetized iron calorimeter (ICAL) detector proposed at the India-based Neutrino Observatory will be a 51 kton detector made up of 151 layers of 56 mm thick soft iron layers with 40 mm air gap in between where the RPCs, the active detectors, will be placed. The main goal of ICAL is to make precision measurements of the neutrino oscillation parameters using the atmospheric neutrinos as source. The charged current interactions of the atmospheric muon neutrinos and anti-neutrinos in the detector produce charged muons. The magnetic field, with a maximum value of $\sim$ 1.5 T in the central region of ICAL, is a critical component since it will be used to distinguish the charges and determine the momentum and direction of these muons. The geometry of the ICAL has been optimized to detect muons in the energy range of 1-15 GeV. It is difficult to measure the magnetic field inside iron, therefore measuring field using external methods can introduce error. In this study the effect of error in measurement of magnetic field in ICAL is studied. An attempt is made to know how the uncertainty in the magnetic field values will propagate in the reconstruction of momentum and other aspects of the physics analysis of the data from ICAL detector using GEANT4 simulations.