This study aims to examine pedestrians self-reported violating crossing Behavior Intentions by applying the theory of planned behavior (TPB).The Behavior Intentions were studied in relation to Instrumental Attitude, Subjective Norm and Perceived Behavioral Control, which were the basic components of TPB.Other factors were also added, including Descriptive Norm, Perceived Risk and Conformity Tendency to investigate specifically their impact on pedestrians Behavior Intentions as an extension of the TPB.Two questionnaires were developed with different scenarios: the first scenario where pedestrians jaywalk in the middle of the block and the second scenario where pedestrians cross the road violating the pedestrian lights at an intersection.Each item in the questionnaire was measured on a 7-point Likert scale.The survey was conducted in the city of Dalian in China, and a total of 499 valid samples were collected, 239 for the first scenario and 260 for the second scenario.Reliability and validity of the data set for each scenario were evaluated.The data was analyzed using the structural equation modeling (SEM), in which four models were built for each scenario and estimated using the maximum likelihood method.For both scenarios, the results show that people has a negative attitude towards pedestrians crossing behavior in a risky way, They also perceived social influences from family and friends, and they have the belief that this kind of risky behavior would lead them to an accident and probably they would get injured or their lives would be endangered.For Scenario 1, results reveal that Instrumental Attitude is the strongest predictor followed by Subjective Norm.Conformity Tendency has shown to be a significant predictor of pedestrians behavior after its addition in the model.Perceived Risk was not found to be a significant predictor indicating that Chinese pedestrians dont perceive the risk in this type of crossing behavior.For Scenario 2 the results also show that Instrumental Attitude is the strongest predictor followed by Descriptive Norms.Once Conformity Tendency is added to the model this construct becomes the strongest predictor showing that the presence of other pedestrians influence behavioral intention.Perceived risk was not significant until the addition of Conformity Tendency.In both scenarios, Perceived Behavioral Control was not significant in any of the models under consideration of the study.In addition to the results, the analysis with regards to the similarities and differences found in two scenarios would help to design a better intervention, such as changing peoples attitude toward this behavior in order to reduce pedestrians causalities or fatalities.Given the welfare implication of planned behavior to predict pedestrians causalities, the Government should vigorously pursue prudent policies conducive to minimize causalities, commit to a developed strategic plan to secure the safety of pedestrians and increase their welfare output. 　　Key Words:pedestrian;violating road-crossing behavior;theory of planned behavior; traffic safety;structural equation modeling (SEM)