The unchecked lateral movement of vehicles across the width of the road is observed on the roads of developing countries like India. This lateral movement can cause distress to drivers as well as passengers and can even lead to crashes on the road. The focus of this study is to analyse the lateral movement behavior of vehicles in mixed traffic conditions and develop suitable prediction models. Data collection was done using videography technique at an urban undivided road stretch identified at Vazhuthacaud, Trivandrum in the state of Kerala, India. Details such as vehicle composition, centre line crossing statistics, whether the subject vehicle is being overtaken, whether the subject vehicle is overtaking a leader vehicle, presence of opposing vehicles, details of the opposing vehicle, details of the leading vehicle, choice of lateral movement, etc. were extracted from the video data. Linear regression based lateral placement prediction model and multinomial logistic regression based lateral movement prediction model were developed. Vehicle Type, Vehicle Speed, Centre Line Crossing, Lane Selection and Leader vehicle Speed were the significant variables identified in the case of lateral placement prediction model. The lateral movement prediction model had Lateral Separation, Lateral Placement and Opposing Vehicle Speed as the significant variables.