Sorry this is a cut and paste answer from https://pharma.elsevier.com/pharma-rd/approaches-for-creating-a-covid-19-vaccine/
Before the coronavirus disease 2019 (COVID-19) pandemic, no one had ever heard of the causative agent 2019 novel coronavirus (SARS-Cov-2). There were no approved antivirals or vaccines for this virus or for any related coronaviruses (e.g., severe acute respiratory syndrome–associated coronavirus (SARS-CoV) or Middle East respiratory syndrome coronavirus).
Problems with COVID-19 vaccine development
Developing vaccines from scratch—which is required for COVID-19—is typically not a good option for stopping an ongoing pandemic. The extensive time required to develop a safe and effective vaccine that can prevent future disease in vaccine recipients (average of approximately 10 years)  means that the public health threat will end long before a candidate vaccine is licensed.
Vaccine development requires extensive planning regarding vaccine design, vaccine production and purification, preclinical testing in animals (to ensure some safety in humans), and multiple phases of clinical trials in humans (phase 1 for safety and phases 2 and 3 for efficacy). Companies taking the risk and proceeding with vaccine development for SARS-CoV-2 (at least 13 companies as of March 7, 2020)  are banking on the virus continuously circulating, so they have a population to conduct phase 2 and 3 trials with. If the virus disappears from circulation in China—or the perceived risk the virus poses to human health substantially declines—before phase 2 trials end, vaccine development will probably end.
Vaccine design approaches
Vaccines that are safe and stimulate the right immune responses that confer protection from disease are challenging to design. There are a number of different strategies companies have used: live attenuated or inactivated viruses, virus-like particles or other protein-based approaches, viral vector–based vaccines or nucleic acid–based vaccines. Of the potential SARS-CoV-2 vaccines in the pipeline, four involve nonreplicating viruses or protein constructs, four have nucleic acid–based designs, two contain live attenuated viruses and one involves a viral vector . Regarding the previous SARS pandemic, one inactivated SARS-CoV  and one DNA-based vaccine  made it through phase 1 trials before vaccine development ended.
The different approaches for vaccine design all carry different advantages and disadvantages . Approaches involving replicating viruses stimulate robust immune responses, but safety is often a concern. Protein vaccines and nucleic acid–based vaccines are often safer but typically have less immunogenicity, and they require adjustments to induce stronger immune responses. Nucleic acid–based vaccines are typically the fastest to get into phase 1 studies, but no nucleic acid vaccine has been licensed for use in humans as of yet.
Because of safety concerns involving older populations (who have higher risk for severe disease), using a live attenuated virus might not be the best approach . Patients with severe disease typically have T-helper 2 responses (immune responses typically induced against extracellular parasites), so vaccines known to elicit this type of immune response (inactivated viruses, virus vectors) might also be best avoided. Vaccines involving protein-based (virus protein subunits, virus-like particles, nanoparticles) or nucleic acid–based (DNA or RNA encoding virus structural proteins) designs, which do not have known safety disadvantages, might be the best approaches for SARS-CoV-2.
Unfortunately, our knowledge of the immune response is not advanced enough for us to accurately predict vaccine safety and efficacy. What we have to do (testing multiple different strategies) is exactly what is being done. Only with results from more extensive research will we know the absolute best approaches for SARS-CoV-2 vaccine development.
Furthermore experts have said that the SARS-CoV-2 virus is heavily glycosalated so that potential immune epitopes are hidden from the immune system.