With accelerated urbanisation, the importance of urban characteristic planning has become increasingly prominent. It has been a focus on how to conduct scientific urban characteristic planning evaluation to ensure high quality and sustainable development of urban characteristic planning. Therefore, this study proposed an urban characteristic planning evaluation index system based on support vector machine (SVM). The process uses hierarchical analysis to determine the relative weight values, calculates fuzzy values for indicators that cannot be directly and objectively evaluated and then introduces SVM for precise processing of interval data. The test results showed that the accuracy of the research method was 0.9760, 0.9600 and 0.9750 when taking the maximum bias value of 0.5 in the three datasets. The optimisation time of the research method for the parameters of the UCI (University of California Irvine) dataset remained within 650–750 ms. The evaluation results of the research method fit well with the expert evaluation results. The above results show that the research method has high data accuracy and operational efficiency in the evaluation of urban characteristic planning, which can make reasonable evaluation of urban characteristic planning. This method can provide reliable reference basis for government decision making.