Estimating capacity in non-lane based mixed traffic streams has always been a challenge due to the haphazard behaviour of vehicles. The present study aims at developing a saturation flow estimation methodology for non-lane-based mixed traffic streams, which is necessary for the capacity estimation and signal design of signalized intersections. The methodology involves formulating a base saturation flow model for mixed traffic streams and adjusting it to the site conditions using various adjustment factors. Base saturation flow is described as a function of width using the concept of unit saturation flow. Further, the traffic characteristics affecting this base saturation flow are developed as adjustment factors. The proposed model is calibrated using field data collected from 16 signalized intersections across 5 cities in India. Here, saturation flow from the field is determined using an optimization technique, which estimates the Passenger Car Unit (PCU) value of vehicles in addition to the saturation flow values. These saturation flow values from the field are used to calibrate the width-based base saturation flow model. Further, adjustment factors for bus blockage, blockage by right-turning vehicles and initial surge are developed considering the discharge behaviour of vehicles in mixed traffic conditions. Agreement of the model with the field-observed saturation flow value indicates the reliability of the developed saturation flow model in mixed traffic streams. The developed model is specifically helpful for engineers and planners dealing with non-lane based mixed traffic conditions wherein the discharge pattern in the field does not comply with the ideal discharge curve.