The car sequencing problem (CSP) concerns a production sequence of different types of cars in the mixed-model assembly line. A hybrid algorithm is proposed to find an assembly sequence of CSP with minimum violations. Firstly, the hybrid algorithm is based on the tabu search and large neighborhood search (TLNS), servicing as the framework. Moreover, two components are incorporated into the hybrid algorithm. One is the parallel constructive heuristic (PCH) that is used to construct a set of initial solutions and find some high quality solutions, and the other is the small neighborhood search (SNS) which is designed to improve the new constructed solutions. The computational results show that the proposed hybrid algorithm (PCH+TLNS+SNS) obtains 100 best known values out of 109 public instances, among these 89 instances get their best known values with 100% success rate. By comparing with the well-known related algorithms, computational results demonstrate the effectiveness, efficiency and robustness of the proposed algorithm.